EUROPEAN COMMISSIONHEALTH & CONSUMER PROTECTION DIRECTORATE-GENERAL

Directorate C - Scientific OpinionsC3 - Management of scientific committees II; scientific co-operation and networks

OPINION OF THE

SC I E N T I F I C C O M M I T T E E O N A N I M A L N U T R I T I O NON THE

DIOXIN CONTAMINATION OF FEEDINGSTUFFSAND THEIR

CONTRIBUTION TO THE CONTAMINATION OF FOODOF ANIMAL ORIGIN

ADOPTED ON 06 NOVEMBER 2000

3

Background

The health significance of human exposure to PCBs and dioxins has been subject ofextensive discussions. The most recent assessment of the risks for human health fromPCBs and dioxins has been performed in 19981, when a WHO consultation group agreedon a tolerable daily intake (TDI) of PCDDs/PCDFs (“dioxins”) and dioxin-like PCBs inthe range of 1 – 4 pg Toxic Equivalents (TEQ)/kg body weight, stressing that the upperrange of the TDI of 4 pg TEQ/kg should be considered as a maximum tolerable intake ona provisional basis and that the ultimate goal is to reduce human intake levels below 1 pgTEQ/kg bw/day.

There are clear indications that the major source of human background exposure is food2

(more than 90 %) with food of animal origin being the predominant source. TheScientific Committee on Food (SCF) was asked recently to advise the Commission on thescientific elements necessary for the establishment of limits and/or alternative measuresaiming at reducing the dietary intake of PCBs and dioxins. The recent cases ofcontamination of feedingstuffs (citrus pulp pellets, oils and fats and kaolinitic clay)highlighted the impact of contaminated feedingstuffs as a source of contamination offood of animal origin with dioxins or PCBs.

In order to complement the SCF evaluation of risks related to human dietary exposure,the Commission requests an evaluation of the contribution of feedingstuffs contaminatedwith dioxins, PCBs and dioxin-like PCBs to the contamination of food of animal origin.

1 VAN LEEUWEN, F.X.R., and YOUNES, M.M. Consultation on assessment of the health risk ofdioxins: re-evaluation of the tolerable daily intake (TDI): Executive summary. Food Additives andContaminants, Vol. 17, no. 4 (April 2000). pp. 223 - 240.

2 FÜRST, P, BECK, H., and THEELEN, R.M.C. (1992)Assessment of human intake of PCDDs and PCDFs from different environmental sources. ToxicSubstances Journal, 12, 133-150.

5

Terms of reference

Considering the above, the Scientific Committee on Animal Nutrition is requested:

(1) to identify feed materials which could be considered as sources ofcontamination of feedingstuffs by dioxins, PCBs and dioxin-like PCBs, andto characterise as far as possible the distribution of levels of contamination ineach of them (e.g. average, upper levels)

(2) to assess the contribution of the different identified feed materials as sourcesof contamination to the contamination of food of animal origin, taking intoaccount both the dietary variations in relation to animal categories, includingfish, and the carry-over rate of dioxins and PCBs (matrix and congenerdependent) in the different products of animal origin.

(3) insofar relevant, to evaluate the impact on animal health of the presence ofdioxins and PCBs in the feedingstuffs and at the levels identified under (1)

(4) to identify the eventual gaps in the available data which need to be filled inorder to allow a complete evaluation.

In view of the fact that the above mentioned WHO consultation group recommended theinclusion of the dioxin-like PCBs in the calculation of TEQs, the Committee is asked todistinguish between dioxins (PCDDs/PCDFs), PCBs and dioxin-like PCBs as far as theavailable data allow, in making its evaluation.

Time limit

In view of a possible legislative action in this area, the Committee is asked to completeits evaluation as soon as possible.

7

Executive summary

The SCAN has been asked to advise the Commission on the sources of contamination offeedingstuffs by polychlorinated dibenzo-p-dioxins and dibenzofurans, collectivelyknown as dioxins, and by polychlorinated biphenyls (PCBs), including dioxin-like PCBs.Advice was also sought on the likely exposure of food producing animals (mammals,birds and fish) to dioxins and PCBs, on the carry-over of these compounds to foodproducts of animal origin and on any impact of dioxins and PCBs present in feedingstuffson animal health.

In reaching its conclusions, the SCAN has taken into consideration all of the availablepublished data and additional information obtained by the European Commission fromMember States and other sources. Account has been taken of the environmental sourcesof pollution resulting in the background contamination of all feed materials and also ofany contamination specifically introduced by production conditions, feed processing andduring the transport and distribution of feed materials and feedingstuffs. Because of thescarcity of reliable data on PCBs, these are not included in this analysis.

As a first task, SCAN prepared a list of typical diets for the main animal species andcategories representative of those in common use within the European Union. Wherepossible, values for upper, lower and mean concentrations of dioxins for each feedmaterial were taken from the available data. In practice, because of the scarcity of dataand difficulty of analysis, some feed materials had to be grouped. The upper values useddo not necessarily represent the highest percentile of the total data, but were morearbitrarily chosen to represent worst-case situations. For example, forages grown inproximity to a source of pollution can have upper values substantially higher than thoseof forages grown elsewhere. Particular attention was paid to soil as a source of dioxincontamination. Soil is deposited onto the aerial surfaces of all plants and is consumed byfree-ranging animals before harvest and, in lesser amounts, by other animals post-harvest.

The total dioxin concentration for each diet was then calculated using the low, mean andhigh values for the dioxin content identified for each ingredient or group of ingredients toidentify the main sources of contamination.

The main SCAN conclusions are:

� Fish meal and fish oil are the most heavily contaminated feed materials with productsof European fish stocks (respective means (for dioxins only) of 1.2 and 4.8ng WHO-TEQ/kg DM) more heavily contaminated than those from South Pacific stock (Chile,Peru, respective means (for dioxins only) 0.14 and 0.61ng WHO-TEQ/kg DM) by afactor of ca eight.

� Animal fat (mean 1ng WHO-TEQ/kg DM) is next in order of dioxins concentration.Values observed depend on the bioaccumulation of dioxins in fatty tissues along thefeed/food chain.

� All other feed materials of plant (roughages, cereals, legume seeds) and animal (milkby-products, meat and bone meal) origin contain mean concentrations of dioxinsaround or below 0.2ng WHO-TEQ/kg DM.

� Roughages present a very wide range of dioxin concentrations depending on location,degree of contamination with soil and exposure to sources of aerial pollution,justifying the use of a worst-case assumption in identifying relatively high mean andupper values.

8

� The limited data available on the contamination of feed materials by dioxin-likePCBs indicates that their inclusion would increase the TEQ value for feed materialsof fish origin by a factor of five and that of other feed materials by a factor of two.

� The contribution of individual feed materials to the dioxin content of the whole dietfor farmed animals depends on the intrinsic degree of contamination and theproportion used in the diet. Greatest concerns arise from the use of fish meal and fishoil of European origin. These are most critical when used in diets for farmed fish andwhere fish meal is incorporated in diets of other food producing animals.

� The consumption of contaminated soil may drastically increase the exposure ofgrazing or free-ranging animals to dioxin. However, the bioavailability of dioxinsadsorbed on mineral or organic soil particles is limited.

� The SCAN stresses that depending on the degree and position of chlorination, theindividual dioxins (congeners) exhibit different transfer rates, and that it is notscientifically correct to calculate transfer from feedingstuffs to products of animalorigin on a TEQ basis only. The exercise must consider the individual congeners.

� With regard to the impact of dioxins on animal health, the SCAN considers that noadverse effects from dioxins would be expected in mammals, birds and fishesexposed to the current levels of background pollution. Toxic consequences would beexpected in animals challenged by severe accidental contamination with dioxins orPCBs.

SCAN recommendations:

The reduction of human exposure to dioxins related to food consumption is important toensure consumer protection. Food of animal origin is a predominant source of exposureof consumers to dioxins. As food contamination is directly related to feed contamination,consistent cross-sectorial actions must be taken to reduce final dioxin impact on humanhealth.

An integrated approach should be followed to reduce the dioxin incidence all along thefood chain i.e. from feed materials to food producing animals then to humans. Takingmeasures on feed materials and feedingstuffs is thus an important step to reduce thedioxin uptake by human.

As far as feed materials are concerned,

• Emphasis should be put on reducing the impact of the most contaminated feedmaterials, e.g. fish meal and fish oil from Europe, on overall diet contamination. Thiscould be achieved by substituting the most contaminated by lesser contaminatedsources, by reducing their intrinsic contamination or by using non (less) contaminatedalternatives, continuing to meet the animal nutrient requirements.

• Any material (recycled products, raw materials, ingredients) used in the manufacturingof feed materials should be guaranteed for quality and safety so that it would notbecome a source of contamination.

• Good manufacturing practices as well as use of Hazard Analysis and Critical ControlPoint principles should be introduced/continued at feed industry level to control thedioxin contamination along the different steps of the manufacture of feed.

• Efforts should be made to reduce dioxins contamination of feed resources at farmlevel (Good agricultural practices) and controlling local conditions of livestock

9

production (e.g. direct environment of dairy farms, free range animal production), inparticular in areas where soil contamination is elevated.

In addition, considering the impact of the environmental pollution on the contaminationof feed materials, measures implemented with the aim to reduce the general dioxinburden should be actively continued.

There is a need for data on dioxins, but also dioxin-like PCBs and PCBs contaminationof the widest range of feed materials and feedingstuffs in order to determine backgroundlevels so as to identify unknown contamination sources. A particular effort should be putinto obtaining more data for feed materials for which a wide range of contamination isreported.

Scientific cooperation should be promoted in order to collect and collate informationavailable in the different Member States at the EU level.

Monitoring programs could be organised at European level in order to widen the currentlimited geographical basis of the information on feed materials contamination.

Regarding the huge amounts of feedingstuffs being produced and sold world wide, feedmaterials imported from third countries should be checked for their dioxin content withthe aim to avoid additional dioxin burden on feedingstuffs.

Data should be obtained from fully identified samples (contaminated or not) andtechniques. In particular, two important aspects: indications or checks, whether sampleswere possibly contaminated, and the definition of the limit of determination have to beconsidered and reported to make data useful and useable. The concept of upper bounddetermination limits should therefore be implemented.

Analytical means should be developed and appropriate analytical requirements adoptedaccording to the aim of the analysis carried out. For instance, in the case of checks ofdioxins levels for control purposes, the analytical limits of determination should be in therange of one fifth of the regulatory limits whereas for control of time trends ofbackground contamination, the limit of determination should be clearly below the meanof the present background ranges for the different matrices.

Emphasis should be put also on quality and qualifications of laboratories involved inmonitoring programs and control activities of or for feed producers. Interlaboratorycalibration studies should be promoted, using reference materials with certified dioxinand dioxin-like PCBs contents which should be made available. With this regard, theCommittee welcomes the recent initiative of the European Commission (2000/C 290/05)3

inviting for submission of proposals to support the development of Certified ReferenceMaterials, in particular in the field of environmental contaminants in food and animalfeed (topic IV.20) and specifically for PCDDs, PCDFs and PCBs in food and feedproducts. Such actions should be promoted.

Basic research is needed on studying the carry-over and establishing pertinent transferfactors for dioxin-like PCBs congeners from soil and feed to animals tissues and products(milk, eggs).

3 OJ n° C 290 of 13.10.2000, p. 4

11

Table of content

1. Introduction..............................................................................................................171.1. Dioxins (polychlorinated dibenzo-p-dioxins and dibenzofurans)...................171.2. Polychlorinated biphenyls...............................................................................171.3. Polybrominated dioxins and furans ................................................................191.4. Tolerable daily intake of dioxins and PCBs....................................................191.5. Human Exposure.............................................................................................20

2. Dioxin sources of feedingstuffs contamination .....................................................222.1. Major cases of dioxin contamination of feed since 1997 ...............................222.2. Environmental contamination of feed materials by dioxins ...........................232.3. Biogenic origin of dioxins in minerals............................................................242.4. Identification of hazards along the production and processing of feed

materials and feedingstuffs .............................................................................252.4.1. Production of feed materials..................................................................252.4.2. Processing of feeds materials and feedingstuffs....................................252.4.3. Collection, transport and storage of feed materials and

feedingstuffs ..........................................................................................262.5. Other contamination sources...........................................................................26

3. Main feed materials used in European food producing animals' diets ..............273.1. Ruminants .......................................................................................................27

3.1.1. Bovines ..................................................................................................293.1.1.1. Extensive rearing ..................................................................293.1.1.2. Semi-intensive rearing ..........................................................293.1.1.3. Intensive rearing....................................................................303.1.1.4. Types of diets ........................................................................313.1.1.5. Veal calf (pre-ruminant stage) ..............................................31

3.1.2. Other ruminants .....................................................................................323.1.3. Soil ingestion aspect ..............................................................................33

3.2. Pigs..................................................................................................................333.3. Poultry.............................................................................................................353.4. Rabbit..............................................................................................................373.5. Fish..................................................................................................................38

4. Contamination levels of feed materials - Data available ......................................394.1. Detection and analysis - Technical considerations .........................................394.2. Evaluation of the contamination of feed materials .........................................41

4.2.1. Vegetable feed materials .......................................................................414.2.1.1. Roughage (forages, conservates and cereal straw)................414.2.1.2. Roots and tubers....................................................................424.2.1.3. Cereals, seeds, by-products of plant origin ...........................424.2.1.4. Vegetable oils........................................................................43

4.2.2. Feed materials of animal origin .............................................................44

12

4.2.2.1. Fish meal...............................................................................444.2.2.2. Fish oil ..................................................................................464.2.2.3. Meat and bone meal, animal fat ............................................474.2.2.4. Milk replacers .......................................................................47

4.2.3. Soil and other feed materials .................................................................484.2.3.1. Soil ........................................................................................484.2.3.2. Binders and anticaking agents...............................................484.2.3.3. Trace elements / Macrominerals / Premixes .........................49

4.2.4. Evaluation of the levels of contamination .............................................494.3. Estimation of background contamination of feed materials ...........................49

4.3.1. Estimation of background contamination on the basis of feedmaterials data.........................................................................................50

4.3.2. Estimation of background contamination of feed materials ofvegetable origin, on the basis of food data ............................................54

4.3.3. Estimation of background contamination of feed materials ofanimal origin, on the basis of food data ................................................55

5. Total contamination estimates of typical diets and identification of themain feed materials contributing to the contamination. ......................................565.1. Ruminants .......................................................................................................575.2. Pigs..................................................................................................................585.3. Poultry.............................................................................................................595.4. Rabbits ............................................................................................................605.5. Fish..................................................................................................................615.6. Identification of the main contributors to diets contamination .......................62

6. Carry over of dioxins from feed to food ................................................................636.1. General findings..............................................................................................636.2. Calculation of transfer factors.........................................................................636.3. Results from relevant investigations...............................................................646.4. Use of TEQ values for carry-over calculation ................................................67

7. Animal health aspects..............................................................................................68

8. Conclusion ................................................................................................................698.1. Identification of contaminated feed materials.................................................698.2. Contribution to the contamination of food of animal origin...........................718.3. Impact on animal health..................................................................................718.4. Need for data...................................................................................................71

9. Recommendations....................................................................................................72

13

10. Bibliographic references .........................................................................................74

11. Annex A: Data available ........................................................................................8111.1. Table A1.................................................8111.2. Table A2.................................................8311.3. Table A3.................................................8311.4. Table A4.................................................8311.5. Table A5.................................................8411.6. Table A6.................................................8511.7. Table A7.................................................8511.8. Table A8.................................................8511.9. Table A9.................................................8611.10. Table A10...............................................8711.11. Table A11...............................................8711.12. Table A12...............................................8811.13. Table A13...............................................8811.14. Table A14...............................................8911.15. Table A15...............................................8911.16. Table A16...............................................9011.17. Table A17...............................................91

12. Annex B: Calculations............................................................................................9212.1. Ruminants .......................................................................................................92

12.1.1. Concentrates Tables B1 & B2......................................9312.1.2. Dairy cattle diets Tables B3 to B6......................................9412.1.3. Beef cattle diets Tables B7 to B10....................................9512.1.4. Calves Tables B11 & B12..................................96

12.2. Pigs..................................................................................................................9712.2.1. Piglets Tables B13 & B14..................................9712.2.2. Pigs for fattening Tables B15 & B16..................................9812.2.3. Sows..................................................................................................99

12.2.3.1. Lactating sows Tables B17 & B18..................................9912.2.3.2. Pregnant sows Tables B19 & B20................................100

12.3. Poultry...........................................................................................................10112.3.1. Broilers Tables B21 & B22................................10112.3.2. Layers Tables B23 & B24................................10212.3.3. Turkeys Tables B25 & B26................................103

12.4. Rabbits Tables B27 & B28................................10412.5. Fish Tables B29 & B30................................105

15

Acknowledgement

SCAN adopted this opinion on the basis of the preparatory work done by the SCANad'hoc Working Group, which included members of the Committee and the followingexternal experts:

• Dr Jacob de Boer

• Dr Peter Fürst

• Dr Rainer Malisch

• Ir Wim Spreeuwenberg

17

1. INTRODUCTION

1.1. Dioxins (polychlorinated dibenzo-p-dioxins and dibenzofurans)

Polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs)belong to the group of lipophilic and persistent organic contaminants.PCDDs and PCDFs are often referred to simply as “dioxins”. Depending onthe degree of chlorination (1 - 8 chlorine atoms) and the substitution pattern,one can distinguish between 75 PCDDs and 135 PCDFs, called "congeners".Although dioxins do not have any benefits and therefore, with the exceptionof research and analytical purposes, were not produced specifically, thesecontaminants have meanwhile found an ubiquitous distribution due to theirformation as unwanted and often unavoidable by-products in a number ofindustrial and thermal processes.

The toxicity of dioxins differs considerably. In particular, those congeners,which are substituted in the 2,3,7,8-position, are of special importance. Thus,of the 210 theoretically possible congeners, only 17 are of toxicologicalconcern. These compounds show a similar toxicity to that of the most toxiccongener 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD), also knownas "Seveso-dioxin".

In order to facilitate the comparison of analytical and exposure data, it hasproven useful in the past to convert the analytical results of the determinationof all 17 congeners of toxicological concern into one summarising resultwhich is expressed as toxic equivalents (TEQ). This conversion is based onthe assumption that all dioxin congeners show similar qualitative effects(binding to the same dioxin-receptor) but with different intensities. Thedifferent binding activity is expressed by toxic equivalency factors (TEF),estimated from the weaker toxicity of the respective congener in relation tothe most toxic compound 2,3,7,8-TCDD, which is assigned the arbitrary TEFof 1. Moreover, it is assumed, that the effects are nor synergistic orantagonistic, but additive. By multiplying the amount of each congenerpresent with the corresponding TEF, the TEQ value of a sample is generated.Different TEF models have been developed in the past. On the internationallevel, the TEF values proposed in 1988 by a NATO/CCMS working groupand the 1997 revised TEFs proposed by a WHO expert group are the mostwidely used (Van den Berg et al., 1998). The resulting total toxic equivalentsare expressed as I-TEQ (NATO/CCMS) or WHO-TEQ (WHO).

1.2. Polychlorinated biphenyls

Polychlorinated biphenyls (PCBs) belong to the group of chlorinatedhydrocarbons, which are synthesised by direct chlorination of biphenyl.Depending on the number of chlorine atoms (1-10) and their position at thetwo rings, there are 209 different compounds, called “congeners”,theoretically possible.

18

In contrast to dioxins, PCBs were produced for technical use. The technicalproducts were liquids with different viscosity depending on the degree ofchlorination (between 42 and 60 % chlorine) and mixtures with dozens ofcongeners. Due to their physical and chemical properties, such as non-flammability, chemical stability, high boiling point, low heat conductivityand high dielectric constants, technical PCB mixtures were widely used in anumber of industrial and commercial open and contained applications.Contained applications include the use of PCBs in hydraulic and heattransfer systems as well as cooling and insulating fluids in transformers andcapacitors. The use of PCBs in pigments, dyes, repellents and carbonlesscopy paper or as plasticizers in paints, sealings, plastics and rubber productsare typical examples of open applications. It is estimated that more than 1million tons of technical PCB mixtures have been produced world-widesince their first commercial use in the late 20s.

Concern over their persistence in the environment and their toxicity has ledto strict regulations within the European Union and other countries of theWestern World in the past two decades. Although the manufacture,processing and distribution of PCBs has been prohibited in almost allindustrial countries since the late 80s, their entry into the environment cannot be excluded, especially due to improper disposal practices or leaks intransformers and hydraulic systems. In this respect, it should also be notedthat, in accordance with Council Directive 96/59/EC of 16 September 1996on the disposal of polychlorinated biphenyls and polychlorinated terphenyls(PCB/PCT)4, the deadline for the decontamination and/or disposal ofequipments and the PCBs contained therein shall be effected at the latest bythe end of 2010.

Technical PCB mixtures always contain a certain amount of polychlorinateddibenzo-p-dioxins and dibenzofurans, formed as unwanted by-productsduring synthesis. This amount can be increased significantly when PCBs areinvolved in fires or are heated in the presence of oxygen at temperaturesbelow 1000°C.

From a toxicological point of view, PCBs can be divided into three groups.While non-ortho and mono-ortho substituted PCBs show highertoxicological properties that are similar to dioxins, and therefore are oftentermed “dioxin-like PCBs”, the di-ortho substituted PCBs are less toxic andhave different toxicological properties. Based on the available toxicologicalinformation, the non-ortho PCB congeners 77, 81, 126 and 169 and themono-ortho PCB congeners 105, 114, 118, 123, 156, 157, 167 and 189 wereassigned a toxic equivalency factor (TEF) by a WHO expert group in 1997(van den Berg et al., 1998). The mono-ortho PCB 28 and the di-ortho PCBs52, 101, 138, 153 and 180 (as well as sometimes also the mono-ortho PCB118) are called “marker PCBs” and are the decisive congeners in regulationsof tolerances for PCBs in all kind of matrices so far.

4 OJ n° L 243 of 24.09.1996, p. 0031 - 0035

19

1.3. Polybrominated dioxins and furans

In addition to the chlorinated compounds, brominated dioxins (PBDDs) andfurans (PBDFs) are of toxicological concern. The Federal Health Office(Bundesgesundheitsamt) in Germany concluded from the available data thatin principle the same toxicological effects were induced. Therefore, itrecommended the application of the same TEFs as used for the chlorinatedcompounds (Appel, 1989, Appel, 1992). If brominated and chlorinatedcompounds are present simultaneously, the sum of both should beconsidered for the recommended tolerable daily intake. However, much lessdata are available for a final conclusion.

The available data indicate that the levels of PBDDs/PBDFs in the foodchain are much lower than of PCDDs/PCDFs. Thus, these compounds werenot considered during the WHO re-evaluation of TEFs in 1997, during thedevelopment of the new TDI recommendation of WHO in 1998 or in thecompilation of EU dioxin exposure and health data in 1999.

PBDDs/PBDFs can derive from the use of flame retardants for fireprotection, such as polybrominated diphenylethers, bromophenols(tetrabromobisphenol A) or polybrominated biphenyls. Laboratory studieshave shown that certain flame retardants and ignition-proof plastics canrelease under thermal exposure large amounts of PBDDs and particularlyPBDFs.

There are no indications that significant amounts of PBDDs/PBDFs haveentered the food chain. However, accidents with fires involving brominatedflame retardants could have had a local effect. It is not clear whetherbrominated compounds have the same stability (e.g. against UV radiation).There is concern that the global use of brominated flame retardants couldlead to increasing levels of PBDDs/PBDFs in the future.

1.4. Tolerable daily intake of dioxins and PCBs

At certain levels of exposure and body burden, dioxins and dioxin-like PCBscan lead to severe perturbations of immune and endocrine functions andreproduction as well as to the development of malignant tumours. Based onnew data on adverse effects of 2,3,7,8-TCDD on reproduction andneurobehaviour in rodents and monkeys, together with novel information onthe molecular and cellular mechanisms of dioxin toxicity, the World HealthOrganisation (WHO) revised in 1998 its recommendation from 1990 andproposed a new tolerable daily intake value (TDI) for humans of 1-4 pgWHO-TEQ/kg body weight (Van Leeuwen and Younes (2000)). Theconsultation stressed that the upper range of the TDI of 4 pg TEQ/kg bwshould be considered a maximal tolerable intake on a provisional basis andthat the ultimate goal is to reduce human intake levels below 1 pg TEQ/kgbw/day. Besides dioxins, this TDI value includes also the above mentioned12 polychlorinated biphenyls (PCB) which show dioxin-like effects.

In the past, most of the toxicological experiments were not performed withindividual congeners, but with technical PCB mixtures. For this reason, it isoften very difficult to attribute observed toxicological properties to

20

individual PCBs or to PCBs as a whole, because it can not always beexcluded that these effects were in fact overlaid for example by dioxinsbeing present as contaminants. Nevertheless, these experiments provided thebasis for the estimation of a tolerable daily intake (TDI) value which wasproposed by OECD in 1988 and adopted in many countries as 1 µg totalPCB/kg body weight. This TDI is about 6 orders of magnitude higher thanthat for dioxins.

1.5. Human Exposure

Humans are exposed to dioxins and PCBs through either

– occupational exposure,

– accidental exposure or

– environmental (background) exposure.

In the past few years a number of studies have been performed in order toestimate the human dioxin exposure and the contribution of the differentroutes to the body burden of these compounds. While specimens fromhumans exposed to background contamination show a similar congenerpattern, those from individuals occupationally or accidentally exposed aredominated mostly by the PCDDs/PCDFs congeners which are characteristicof the respective sources of contamination

Human dioxin exposure through background contamination is possible byseveral routes:

– inhalation of air and ingestion of particles from air,

– ingestion of contaminated soil,

– dermal absorption,

– food consumption.

The first three routes normally contribute less than 10% to the total dailydioxin intake, while more than 90% of human dioxin exposure derives fromfood. About 90% of the latter come from food of animal origin. As is seenfor humans, the dioxin body burden of animals (and therefore thecontamination of food products of animal origin) originates mainly fromfeed contamination. Therefore feedingstuffs are of special concern in thecarry over of dioxin contamination to the human food chain.

Feed, food producing animals and finally food products of animal origin maybecome contaminated with dioxins through e.g.:– deposition of emissions from various sources on farmland,– burning of raw material containing potential dioxin sources for direct

drying,– blending of feedingstuffs with dioxin containing products and/or raw

materials,– application of contaminated pesticides, detergents, disinfectants etc.,– contact with / consumption of wooden materials treated with wood

preservatives,– application of sewage sludge on fields,

21

– flooding of pastures,– contamination of water with waste water and effluents,– food processing– migration from chlorine bleached packaging material.

Due to the lipophilic properties of dioxins and their well-knownaccumulation in the food chain, food samples of animal origin are of specialimportance for a risk assessment of human exposure. While normallyspecimens of plant origin normally only contain levels near the detectionlimit, all samples of animal origin show typical patterns of PCDDs andPCDFs. Predominant congeners are those with 2,3,7,8-chlorine substitutionbelonging to the toxic dioxin compounds.

At the end of the 80s, for reasons of safety, various countries took measuresto reduce the dioxin emissions and to lower their levels in the environment,because at that time it was estimated that the daily dioxin intake exceededthe TDI value proposed by several health authorities. The success of thesemeasures is demonstrated by a significant reduction of the dioxincontamination of the environment and particularly of food in many countries.For Germany, The Netherlands, and the United Kingdom, data are availablefor a number of years, allowing time trend analysis (European Commission,1999a). In The Netherlands a constant decrease of 50% of the intake of I-TEQ/kg bw/day over each interval of 5.5 years over the period 1978 to 1994was observed. In the United Kingdom, three data points are currentlyavailable, for 1982, 1988 and 1992, that show the exposure has also fallenfrom 4 pg I-TEQ/kg bw/day in 1982 to 2.1 in 1988 (48% decrease) and 1.15in 1992 (another 45% decrease). Recent investigations of German foodsamples have shown dioxin levels approximately 50% lower thancorresponding samples collected and analysed 10 years ago. As a result, theaverage daily dioxin intake for adults was reduced in the meantime fromaround 2 pg TEQ/kg body weight to below 1 pg TEQ/kg body weight.However these very consistent results show that even in those countrieswhich took measures to minimise dioxin emissions, the average daily dioxinintake is in the range of, or still exceeds, the TDI value proposed by WHO in1998. If dioxin-like PCBs are also considered, the resulting intake isapproximately doubled and the present picture worsened.

Where monitoring programmes are available, important reductions in thedioxin content of human blood and mothers' milk could be observed in thepast ten years. However, data from Germany and Spain indicate that thisreduction seems to be levelling out at the end of the 90s. The successfulreduction of the dioxin content in man at the beginning of the 90s is theresult of measures to reduce emissions from various environmental sources.At the end of the 90s, however, more focus has been put on the possiblecontribution of feedingstuffs to exposure of human beings to dioxins, as aconsequence of a more intensive monitoring, which followed some episodesof critical contamination of feedingstuffs by dioxins.

Comparable data on PCBs are still scarce and often are difficult to comparebecause the selection of PCBs determined in the different studies variesconsiderably. Limited data available from industrialised European countriesindicate that the average daily total PCB intake for humans through

22

background exposure is approximately one tenth of the proposed TDI value.The assessment of dietary intake of dioxins and dioxin like PCBs by thepopulation of EU Member States (SCOOP task 3.2.5, June 2000) revealedlarge differences in the amount, detail and quality of the data from theparticipating countries. Therefore, the participants of the SCOOP taskrecommended that the collection of occurrence and food consumption datashould be repeated every 5 to 10 years because it was considered extremelyvaluable to compare and follow temporal trends in the exposure of thepopulations of different European countries to dioxins and relatedcompounds.

2. DIOXIN SOURCES OF FEEDINGSTUFFS CONTAMINATION

2.1. Major cases of dioxin contamination of feed since 1997

The finding of a high dioxin contamination in feed material for fish andpoultry in the USA in 1997, in citrus pulp pellets (CPP) from Brazil in 1998or in fat from Belgium 1999 raised the question whether certain feedingstuffsbear a higher risk of a possible dioxin contamination. Thus, the major“accidents” are described and discussed in more detail:

(1) In 1997, the FDA identified the source that caused elevated levels ofdioxin in two poultry samples as that of a clay material called “ballclay” used as anti-caking agent. Feed manufacturers and commercialcatfish and egg producers were warned not to ship products that maybe contaminated. This was the first finding of a dioxin contaminationof this type of technological additive. For a long time, the reason ofthe contamination remained unclear. In 1999, a similar contaminationoccurred in Europe in kaolinitic clay which was used as “anticakingagent” as well, and for production of mineral feed. Gradually itbecame obvious that a natural source was involved. Possibly,geothermal processes formed this unique pattern of dioxins over timefrom organic material and chlorine. Data presented at the dioxincongress in Venice in September 1999 showed that this unique dioxinpattern could be found in different regions in the world: in theMississippi basin (where the ball clay was mined), in the Westerwaldarea in Germany (where the kaolinitic clay was mined) and insediments at the East coast of Australia.

(2) In 1998, citrus pulp pellets (CPP) from Brazil with high dioxincontamination were found (Malisch, 2000). Comprehensiveinvestigations revealed that the use of highly contaminated lime(calcium hydroxide) was the source of the dioxin contamination ofthis CPP. Lime is added to wet peels, seeds and pulps of oranges inorder to facilitate the drying process and to raise the pH from between2 and 3 up to between 6 and 7. Lime constitutes about 2 % of thedried CPP. In the past, the CPP producing industry in Brazilpurchased quicklime or hydrated lime either from suppliers of mined“virgin” lime or from lime converters. The highly contaminated limeused for CPP production came from one supplier. It turned out thatthis supplier was a converter. This lime converter purchased, at that

23

time, one of the main ingredients, lime milk, from a specific limemilk supplier who generated the lime milk as a by-product from aproduction process. Thus, citrus pulp as such should not beconsidered as a dioxin risk but the use of chemical by-products orwaste for manufacturing of feed materials, which may lead to theircontamination.

(3) In 1999, in Belgium the contamination of fat used for production offeedingstuffs caused a severe contamination of different animalproducts. Finally it turned out that the discharge of a technical PCBmixture at fat collection sites used for feedingstuffs production hadcaused this massive dioxin contamination. Used PCB products arewell known for their high dioxin contamination and have to bedischarged as hazardous waste.

(4) In 1999, grass meal with high dioxin contamination was found in theGerman state Brandenburg. Here, the dioxin contamination camefrom the drying process in which all types of wood were burnt,including waste wood with chemical contamination from paints orpreservatives.

(5) In June 2000, dioxin levels were found in certain premixturescontaining choline chloride originating from Spain. Classified as aprovitamin, choline chloride is used as an animal feed additive.Investigations tracing back the source of contamination revealed thatthe pure choline chloride was not the source of the problem but thecarrier of plant origin was contaminated. Although declared as corncob meal, the carrier also included rice husks and/or saw dustpresumably treated with a wood preservative. The congener patternfound in the contaminated lots matched those typical ofpentachlorophenol contamination, which is used as a woodpreservative.

All the above cases demonstrate that contamination of feedingstuffs bydioxins and PCBs may occur at different levels of the feed chain and havevery different origins:

• Additives from either natural or synthetic origin used in feed processingas a possible dioxin source (e.g. kaolin and lime)

• Technological processes applied to feed, directly generatingcontamination (e.g. drying process)

As a result, the inventory of dioxin sources which was used to reduce thedioxin emissions into the environment should also be extended to thedomain of feed materials and feedingstuffs production.

2.2. Environmental contamination of feed materials by dioxins

The contamination of the environment by dioxins is primarily caused by theaerial distribution and deposition of emissions from various sources (wasteincineration, production of chemicals, traffic, etc.). Moreover,application/disposal of chemicals could contribute to more severe and

24

localised contamination of the environment in general and of feed materialsin particular.

Soil is a natural sink for this kind of persistent and lipophilic compounds.Adsorption to the organic carbon fraction of the soil takes place, and onceadsorbed the dioxins remain relatively immobile. Soil is a typicalaccumulating matrix with a long memory, which reflects the base linecontamination of a region. Apart from atmospheric deposition, soils may bepolluted by the application of sewage sludge or composts, and by spills anderosion from nearby contaminated areas. The intake of soil by free-rangegrazing (ruminants, birds) or grazing/burrowing animals (pig, wild boar) isdifferent to a large extent, and take place directly or through dust deposits onvegetables.

As soil to plant transfer of dioxins is very limited (see in the followingchapters), aerial distribution and deposition of dioxins and dioxin-like PCBsare the main sources of contamination of leafy vegetables. Furtherdistribution in the plant is limited and restricts contamination of seeds and ofmost plant by-products. Leaves are either directly grazed by free-ranginganimals, or cropped and then preserved under dried form (hay) or silage.Spreading of sewage sludge, adhering on the vegetation, increases to alimited extent the exposure of livestock (European Commission, 1999a).

Dioxins and dioxin-like PCBs are poorly soluble in water. However they areadsorbed onto mineral or organic particles in suspension in water. Thesurface of oceans and seas is exposed to aerial distribution of thesecompounds and consequently they are concentrated along the aquatic foodchain. Fishes (products) used as feed materials are therefore of particularconcern. This is especially true for those areas where in addition wastewaters or contaminated effluents from certain processes, such as paper pulpbleaching, are introduced into the aquatic system. Terrestrial free rangedomestic animals and semi-wild animals drinking river or pond water can beexposed also to a limited extent.

2.3. Biogenic origin of dioxins in minerals

The finding of uncommon but similar PCDDs and almost PCDF-freeprofiles in different types of clays protected from the present anthropogeniccontamination, strongly suggests a natural origin for these compounds (JobstH., 2000). Comparable patterns for the most abundant hexa-, hepta- andocta-CDDs were found in sewage sludge both from 1933 and 1981/82,despite the fact that production of pentachlorophenol (PCP) did notcommence until 1938. This suggests that a natural chlorination process isinvolved (Lamparski et al., 1984). More recently, it has been shownexperimentally (Oberg et al., 1992; Rappe et al., 1999) that normal sewagesludge fortified with isotopically labelled PCP generates after a few weekshepta- and octa-CDDs (but not penta- and tetra-CDDs) through adimerization reaction. Important natural sources of PCP such as compostinghave been reported (Sievers et al., 1994; Oberg et al., 1994). Therefore,depending on the very specific local conditions required for thebioproduction of PCP (or other simple organic chlorine compounds) andsubsequent chemical or biochemical dimerization processes that prevailed

25

when the sediments were formed, the deposits exploited today may be eitherlow or highly contaminated with these specific PCDDs.

2.4. Identification of hazards along the production and processing of feedmaterials and feedingstuffs

2.4.1. Production of feed materials

Many by-products of food processing (e.g. starch, sugar, oils and fats, fruitand vegetables transformations) and biotechnology (e.g. biosynthesis oforganic acids and amino-acids) are used as feed materials. The dioxin anddioxin-like PCBs burden of the usual feed materials of plant origin used asraw material is examined further on in the opinion. However, specialattention should be given to a possible contamination, which may occur atcertain steps of the production process of these by-products, namely whenchemical substances like catalysts, solvents, pelleting aids, pH modifiers orfiltration agents are introduced. This is illustrated by the example of thedioxin contamination of citrus pulp.

2.4.2. Processing of feeds materials and feedingstuffs

Feed materials and feedingstuffs undergo usually one or more processingsteps.

a) Mechanical steps

Grinding, mixing, pelleting and extrusion are basic operations in thepreparation of feedingstuffs. These mechanical operations involve frictionshear of different intensities which, even for the most severe (extrusion),may increase the temperature to a maximum of about 230°C. Addition ofsteam is frequently used which does not increase this temperature range, andmay well contribute to reduced friction. No production of dioxins can beexpected to occur at these temperatures.

b) Heating operations

Roasting and micronising operations involve temperatures until 180°C that isstill not high enough to generate dioxins. Only the source of the heat may beof concern. Distillation of fats, oils and fatty acids, as a purification process,is conducted under reduced atmospheric pressure such that the highesttemperatures reached are about 200 to 220°C which is not enough togenerate dioxins.

Drying of feed materials such as green forage, sugar beet pulp or citrus pulp,potatoes or grains, may involve atmospheric temperature air flow or hot airgenerated by a non-polluting source, i.e. electric heating or heat exchange.Under these circumstances no dioxin contamination can be expected.However, other drying methods involving a direct contact between feedmaterials and an air flow heated by a direct combustion process and carryingcombustion products (gases, smoke) may constitute a considerable pollutionsource highly dependent on the nature of the fuel used. Whereas natural gasis considered as a clean energy source, other sources (i.e. oil and derivates

26

including additives, pit-coal, wood), may generate dioxins during thecombustion process, especially if combustion is uncomplete. The particularcase of the Anaerobic Pasteurising Conditioning (APC-treatment), whichinvolves heated steam with the presence of combustion gases, should becarefully considered in this regard.

c) Physico-chemical steps

Chemical treatments or reactions (acidic, alkaline, oxidative) applied tocertain feed materials to improve their nutritional value or othercharacteristics, do not use or generate temperatures above 120°C.

Extraction of oil from oilseeds, palm kernels or coconut products in the oilindustry, but also of animal fat from meat and bone meals in the renderingindustry, implies the use of organic solvents. The presence of dioxins assolvent contaminants, but also the eventual genesis of these compounds fromchemical reactions occurring between the solvent (i.e. organochlorine) andfeed materials, may contribute to the contamination of feedingstuffs.

2.4.3. Collection, transport and storage of feed materials and feedingstuffs

Contamination can occur during collection of feed materials from the fieldsor industrial sites, as well as when open containers are used for selectedwaste deposits (e.g. polluted waste cooking oil and the dioxin crisis inBelgium). Loading and unloading operations are critical steps wherecontamination can occur. In addition, generally, transport is not operated indedicated specific means. For instance, tanks used for industrial (non food /non feed) products can be utilised to transport fat and oils for feed use. Useof paints containing PCBs in containers for transport and storage representan additional risk for the products.

2.5. Other contamination sources

Any direct contact of feed materials or animals to materials, equipment orlitters (sawdust can be used in the beddings of pigs and cows) made of woodtreated (preserved or coated) with chemicals that may contain dioxin(pentachlorophenol), constitutes a potential source of contamination.Accidental pollution of feed materials by dioxins may occur from thelocalised spread out into the agricultural environment of PCBs normally usedas heat exchangers but recycled as lubricants, but also from the uncontrolledcombustion of wasted plastic/rubber materials (chlorinated compounds).

27

3. MAIN FEED MATERIALS USED IN EUROPEAN FOOD PRODUCING ANIMALS' DIETS

The following describes the various EU production systems (standardised intensive,extensive, quality certified or organic), for each of the animal categories identified.Although the diets described here represent classical and usual diets for Europeanproduction of different food producing animals, there exists a wide range of otherpossible feed materials used in Europe as well as in third countries that represent alimited percentage of the ration and correspond to local and circumstantial supplyingconditions.

Unfortunately, the existing data do not allow consideration of the contribution ofindividual ingredients to the total dioxin content of the diet. For this reason, dietarycomponents are specified only in terms of groups of feed materials for whichsufficient data are available.

From a practical point of view, feed materials have been grouped into three maincategories. Each category has been further divided into a limited number of sub-groups (for instance, data concerning wheat, barley and corn appear under“vegetable feed materials” and “cereals”):

A. Vegetable feed materials1. Roughages2. Roots and tubers3. Cereals and seeds4. By-products of plant origin5. Vegetable oils

B. Feed materials of animal origin1. Fish oil, fish meal2. Animals fat, meat and bone meal3. Milk products

C. Other feed materials1. Binders and anticaking agents2. Trace elements3. Macrominerals

This classification will be used in the tables all along this report.

3.1. Ruminants

Feed materials and consumption are highly dependent on resources andrearing conditions.

Ruminant rearing is usually linked to regional characteristics and particularlyto availability of forages. For this reason, rearing practice in the EU countriesare very different, depending on climate (which affects forage availability),intensification level, type of production and breed. Wet temperate zones butalso cold or warm semi-arid temperate areas are encountered withinEuropean Union. Consequently, according to the type of livestockproduction, the theoretical intake of forages and concentrates under different

28

feed regimes can vary considerably and for this reason, different scenarioshave to be considered. They are summarised in table 1.

Table 1 Different rearing scenarios for ruminants and their respective dietcomposition

Diet composition (in % dry matter)SCENARIO

Roughage** Concentrate

Extensive rearing 80-100 0-20*

Semi-intensive rearing 65-90 10-35*

Intensive rearing (20***) 40-80 20-60 (80***)* in winter the intake of concentrate is higher than in summer due to the reduced

availability of forages.** mainly pasture, hay and silage*** beef cattle in feedlots

The kind of roughage fed depends on availability (area, season).

The level of concentrates in the diet is a function of animal performance(milk yield/weight gain) and is aimed to fulfill the energy and proteinrequirements of the animals. The concentrates are compound feedingstuffsmade from different feed materials.

The composition of concentrates is affected by the basal diet, feed costs andfulfilment of technological requirements. Some examples of concentratecomposition are given in table 2.

Table 2 Examples of concentrate composition, for ruminants

Concentrates N°I II III IVFeed materials

PercentageA3 Cereals and legumes 26 55 64 62

Soybean meal 10 14 15 15A4 By products of plant

origin Other by products 60 25 10 10B1 Fish meal - - 5 5B2 Fat* - 2 2 4C Premix** 4 4 4 4

* fat can include fat of animal origin and/or of vegetable origin. However for furthercalculation, it will be considered as exclusively made of animal fat.

** includes minerals, trace elements, vitamins and other feed additives

29

3.1.1. Bovines

3.1.1.1. Extensive rearing

The animals are kept on pasture for all the year (or at least 4 to 8 months).They are housed in the cold periods when the grass production is notsufficient. The main feed source in summer (often the only source) ispasture, which can be turned or free. During lactation, complementary feedcan be given to the animals (premixes and energetic sources, cereals,molasses and proteins).

(1) Milk and meat production

During the summer period, animals are kept outside and the diet istherefore mainly based on fresh grass (90 to 100% of the diet)whereas concentrate can be given (but not always) at up to 10% ofthe diet.

During the winter period, animals are housed for about 4 to 8 months.Fresh grass is replaced by grass hay and silages representing 80 to90% of the diet. The use of concentrate increases up to 10 to 20% ofthe diet. The consumption of concentrate is additionally stronglyincreased when forage availability (hay and silages) is low on thefarm or when daily milk production is increasing. For this reason theintake of mixed feeds in the housing period may vary greatly.

Beef cattle are predominantly pasture-fed in summer, withconcentrates fed in early spring and late autumn.

(2) Nursing cows

The availability of forages due to weather conditions strongly affectsthe feed regimen for nursing cows, which Dry Matter Intake (DMI)ranges from 10 to 12 kg.

In summer the typical diet is based on pasture partially with varyingamounts of hay or straw and mineral vitamin supplement.

In winter the diet is based mainly on silage, partially with loweramounts of hay. The animals receive up to 2 kg of concentrate.

3.1.1.2. Semi-intensive rearing

The animals are housed from the late autumn to the spring season

(1) Milk production

Typical of humid and alpine temperate zones, the animals' diet duringthe summer season is based on pasture, but supplementation withconcentrate (25-30% of DMI) is common practice. During the winterperiod animals are housed for about 5 to 8 months with diet based onsilage, partially also hay and concentrate (30-35% of DMI). The

30

percentage of concentrate inclusion in the diet is positively related tomilk yield.

(2) Meat production

When pasture productivity is high and grass has a high nutritivevalue, cattle diets can be composed almost entirely of fresh grass andconsequently the intake of concentrate is low (around 1 kg/head/dayfor animals between 6 to 12 months of age) or limited to self feedingof minerals and liquid feed (urea) based on molasses.

In summer and particularly in autumn the production of pasturedecreases and an energy supplementation is required. Molasses basedconcentrates or cereals can be given to the animals during thefinishing period (last 3 months). The amount of concentrate rangesfrom 2 to 3 kg/head/day, according to the fattening status of theanimals. During winter time the animals eat silages and hay withlower supplementation of concentrate.

3.1.1.3. Intensive rearing

(1) Milk production

Animals can be reared in fixed stalling or in free stalling. Roughageconstitutes of grass or legumes (alfalfa, fresh, as silage or hay) orcorn silage. In the most productive areas of Europe, pasture is alsoavailable in summer. Forages are usually produced in the farm, butsometimes they can be bought from other places. Silages and hayaccount for 50-80% of DMI, the remaining part being constituted bymixed feeds (concentrates up to 60%).

(2) Meat production

In Europe, diets are based mainly on silages (corn, grass) with lowlevels of straw or hay, and an energy and protein supplementation.During the fattening phase, the dry matter intake is about 2.0 % oflive weight.

In some areas, a special type of beef cattle production, called “babybeef” system, exists. In this case, the diet is based on cereals (mainlybarley). Only a minimum daily supply of fibrous feed is necessaryand no green forage, hay or corn silage are used in the diet.

Beef cattle can also be fed in “feedlots”. In this case, the mainingredients of this diet are cereals, by-products of cereals andsoybean, given in concentrates representing about 80% of the diet.

31

3.1.1.4. Types of diets

All the feeding situations described above for bovines can be summarised inthe tables 3 and 4:

Table 3 Diet composition for dairy cows, expressed in percentage on a drymatter basis

Milk yield (kg/day)

5-10(and

dry cow)

5-10(and

dry cow)15-25 25-35 ≥35

Diet composition in percentageFeed materials

n° 1 n° 2 n° 3 n° 4 n° 5

A1 Roughage 100 90 75 60 40Concentrate 0 10 25 40 60

Dry Matter Intake (kg/day)9-12 9-12 15-18 20-22 ≈25

Table 4 Diet composition for beef cattle, expressed in percentage on a drymatter basis

Average daily gain (g/day)400-800 800-1200 >1200 Feedlots

Diet composition in percentageFeed materials

n° 6 n° 7 n° 8 n° 9

Corn silage - 35 70 -A1 Roughage

Others 80 40 - 20Concentrate 20 25 30 80

3.1.1.5. Veal calf (pre-ruminant stage)

The diets for calves are based on milk replacers.

Milk replacers are made by adding animal and vegetable fat (200g/kgdry matter) to skimmed milk. It is, however, possible to feed calveswith "milkless" powder in which skimmed milk is replaced by wheyproteins, caseins, soybean by-products. A minimum daily supply offibrous feed is necessary for calves over two weeks of age, and thequantity is raised from 50 g to 250 g/day from the 8th to 20th week of

32

life (Commission Decision 97/182/EC5 amending the Annex toCouncil Directive 91/629/EEC6 laying down minimum standards forthe protection of calves). The amount of milk replacers ingestedincreases from 550-700 g dry matter per day in the first week until3000 g dry matter per day at the end of the fattening period.

The total diet consists on a average of 90% milk replacer and 10%roughage (diet n°10, used for calculation in tables B11 and B12).

3.1.2. Other ruminants

The previous described scenarios can fit also for sheep and goats. For smallruminants the DMI as forages can be relatively higher than for bovine.

The meat sheep systems can be categorised as either lowland, which areintensive and involve housing in winter, or hill-farming systems, which aremuch more extensive. Lowland systems are pasture-based in summer, butextensive use is made of supplements such as hay, silage, straw, turnips andconcentrate pellets. The most intensive form is for early lamb production, inwhich ewes lamb indoors and are fed generously to maximise milkproduction, with lambs creep-fed on concentrates, early weaned and fattenedindoors on concentrate-based diets. Hill systems are much more extensiveand based exclusively on forage available on rough, hill grazing for ewe andlamb, possibly with concentrate, hay and turnip supplementation duringtimes of forage shortage. For sheeps, usually grown on pasture, the dataretained for extensively reared animals can be used, with an estimated DMIaround 3-4% of live body weight.

The dairy sheep breeding can be considered a semi intensive system with thelargest part of animals lambing in autumn (except 30-40% in late winter).Grazing is the major forage source except 2-3 months in winter when hayand some silage are used; concentrate range from 20 to 40% of DMIaccording to forage availability and milk yield. The DMI can represent asmuch as 5% of body weight. For 3-4 months of dry season (summer) theanimals are dry and fed crops by-products.

For riverine buffaloes the same conditions as those described for extensive,semi-intensive reared cattle can be used, assuming a DMI by adults animalsof 14-16 kg in relation to the milk production. Nevertheless many buffaloherds are at present reared in a similar way to intensive dairy cows systemi.e. large of use of corn silage (40% DMI) hay and concentrate (30% DMIeach). It must be noted that the soil intake through drinking water fromponds or rivers and through grazing may be considerable but only for theextensive system.

5 OJ n° L 076 of 24.02.1997, p. 30

6 OJ n° L 340 of 11.12.1991, p. 28

33

3.1.3. Soil ingestion aspect

Soil ingestion depends on the abundance and quality of pasture and animals'density at pasture. If grass is vigorous, the animals eat the highest foliaceousparts and the soil ingestion during grazing is low. If pastures are poor (orcattle density too high), the animals graze also the parts near the ground andingest greater amount of soil. The Danish Veterinary and FoodAdministration (Danish Plant Directorate), estimated a daily soil intake of0.225 kg for cows with 18,225 kg of dry matter intake (DMI) (1,23%).According to Fries et al. (1982) soil ingestion can represent from 0,14% upto 2.40% of dry matter, and in worst conditions a bovine of 600 kg bw wouldingest 480 g of soil/day. Mayland (1975) reported, for young bulls, a caselimit of soil intake of 18% of DMI.

Corn for silage, after being cut, is not spread onto soil like grass beforeensiling or sun-curing, but it is immediately machine chopped and processed.For this reason soil contamination of corn silage is lower than that of grasshay or ensiled grass. The use of artificially dehydrated forages as substitutefor sun-cured hay also can reduce the level of dioxin contamination derivedfrom soil, if the combustion process does not emit dioxins.

For riverine buffaloes the intake of soil in suspension in drinking water orthrough grazing could be considerable, but reliable data are not available.

3.2. Pigs

Most of the EU pig production is intensive. Only about 1.5% of sows, usedfor reproduction purposes, are raised outdoor, in particular in the UnitedKingdom and in France (Le Denmat et al., 1995). The following pigcategories/diets have been identified:

(1) Piglets up to 1 month of age are generally suckled by their motherand received very few dry feeds which amounts to a maximum of 1kg creep feed for the whole month. This diet is based on cereals andprotein-rich ingredients including dried skimmed milk (2-10%), driedwhey (2-15%), refatted dairy products (2-10%). Early weaned pigletsfrom 3-4 days of age could be fed diets using higher proportions ofmilk substitutes. The younger the piglets are, the higher is theproportion of milk products in the diets

(2) The young animal diets may include the following feed materials:cereals, soybean meal, fat (up to 8%), fish meal (up to 5%) andminerals. Meat and bone meal can replace soybean meal up to 2 to5%. Manioc can also be used as a cereal substitute up to a level of20%. Examples of diets for piglets up to 2 months of age are given intable 5.

34

Table 5 Examples of diets for piglets aged up to two months in percentage.

Diet composition(in percentage)Feed materials1 2 3

A3 Cereals Barley, Maize, Wheat 55 59 60Soybean meal 25 24 21

A4 By-products of plant originOther by-products 11 - 12

B1 Fish meal 5 5 2B2 Fat* - 8 2C Premix ** 4 4 3

* fat can include fat of animal origin and/or of vegetable origin. However for further calculation,it will be considered as exclusively made of animal fat.

** includes minerals, trace elements, vitamins and other feed additives

(3) For fattening pigs (growing or finishing) from two months untilslaughter, the diet includes cereals, wheat bran, corn gluten feed,molasses and peas, meat and bone meal, animal fat, soybean meal,and other feed materials including minerals. Manioc can be used, asfor piglets, as energy source to substitute cereals up to a level of 40%in the diet of growing pigs. In some particular cases, fatteninganimals receive large amounts of liquid whey and/or fresh dairy by-products. In case of the production of heavy pigs, up to 170 kg liveweight, higher amounts of fibrous materials can be introduced,supplied by by-products of cereals, seeds and sugar.

Some variations of a pig diet are given in table 6.

Table 6 Some typical diets for fattening pigs and gilts in percentage

Diet composition (in percentage)Feed materials 4 5 6 7 8Barley, Maize, Wheat 74 40 65 54 59

A3 Cereals andlegumes Peas - - 15 15 2

Soybean meal 23 23 14 20 20A4 By-products of

plant origin Other by-products - 30 3 3 9Meat and bone meal - - - - 5

B2Fat* - 4 - 5 4

C Premix ** 3 3 3 3 1* fat can include fat of animal origin and/or of vegetable origin. However for further

calculation, it will be considered as exclusively made of animal fat.** includes minerals, trace elements, vitamins and other feed additives

(4) Diets of pregnant and lactating sows are based on the same feedmaterials as for fattening pigs. However, they can contain fibrouscomponents such as citrus pulp (0-15 %) dried sugar beet pulp (0-15

35

%) and alfalfa meal (up to 32 %). Manioc can be used as an energysource up to 40%.

Additionally, for animals kept outdoor, which are mainly productivegilts and sows (pregnant and lactating), the diet includes substantialintake of green forage (2 to 5 kg/day or up to 1 kg DM) and soil (200to 500 g/day) during rooting.

Six examples for lactating sows and pregnant sows are presented intable 7.

Table 7 Examples of diets for lactating sows (n°9, 10 & 11) and pregnantsows (n°12, 13 & 14) expressed in percentage

Diet composition(in percentage)Feed materials

9 10 11 12* 13* 14A1 Roughage - - - - - 32

Barley, Maize, Wheat 63 44 47 42 17 25A3 Cereals and

legumes Peas 16 14 8 12 12 -Soybean meal 15 19 20 3 11 10

A4 By-products ofplant origin Other by-products - 12 19 40 53 30

B1 Fish meal 2 2 - - - -B2 Fat ** - 5 3 - 5 1C Premix *** 4 4 3 3 2 2

* animals fed restrictively** fat can include fat of animal origin and/or of vegetable origin. However for further

calculation, it will be considered as exclusively made of animal fat.*** includes minerals, trace elements, vitamins and other feed additives

3.3. Poultry

The composition of poultry diets is close to that of pigs. Diets of youngpoultry (growing and fattening broilers, turkeys and other poultry) are similarto those of piglets; those of adult poultry (layers, poultry for reproduction)are like those of adult pigs.

Cereals, legumes and by-products (such as extracted oil meals), cerealmiddlings, corn gluten feed etc. are the most frequent feed materials. Fats,manioc, some feeds of animal origin and additives vary according to animaldiets. The proportion of different ingredients varies according to animalrequirements along the growing period. Table 8 gives three typical diets forbroilers, layers and turkeys. Similar diets are fed to ducks, geese, pheasants,guinea fowl and other poultry.

36

Table 8 Typical diets for food producing poultry (broiler (diets n° 1, 2 & 3), layer (diets n°4, 5 & 6), turkey (diets n° 7, 8 & 9)) expressed in percentage

Diet composition (in percentage)Feed materials

1 1) 2 2) 3 2) 4 5 6 7 3) 8 4) 9 5)

A1 Roughage - - - 3 3 5 - - -Cereals 58 70 50 70 60 50 47 60 50

A3 Cereals and seeds(legumes) Legumes (peas, beans) - - 5 - - 10 - - 5

Soybean meal 30 18 21 10 12 8 40 30 20A4 By-products of

plant origin Other by-products - - 10 5 14 16 5 - 14B1 Fish meal 5 - - 3 - - 3 - -

Meat and bone meal - 3 - - 3 - - 3 -B2

Fat* 3 5 10 - - 2 2 4 8Limestone - - - 7 6 7 - - -

CPremix** 4 4 4 2 2 2 3 3 3

1) Starter period (0-2 weeks); 2) Grower period (3-5 weeks); 3) Starter period (0-4 weeks);4) Grower period (5-8 weeks of age); 5) Fattening period (> 8 weeks of age)

* fat can include fat of animal origin and/or of vegetable origin. However for furthercalculation, it will be considered as exclusively made of animal fat.

** includes minerals, trace elements, vitamins and other feed additives

Typical EU-production conditions are characterised by indoor keeping. Insome cases (layers, ducks, geese etc.) free-range production increased. Theproportion of free ranged poultry depends on poultry species and categoriesas well as the country varying between less than 5% (broilers, turkeys) andmore than 50% (ducks, geese) of the total poultry population.

Free ranging of poultry are typical production conditions of organic farming.It is not allowed to feed feeds of animal origin (e.g. meat and bone meal),extracted oil meals, technical amino acids, most of the feed additives andfurther feeds in organic farming systems. Furthermore free ranging poultrymay consume forages, insects and soil. The daily forage intake of layersamounts to about 35 g (legumes, grass, herbs etc., Mehner 1968.). About 20gof feed comes from animal origin (insects, worms etc., Krax 1974, Marks1985). Free ranging hens consume about 4g protein of animal origin per day(Krax 1974).

Information of soil intake of free ranging poultry varies between 2 and 10 gper laying hen and day (Mc Kone 1994, Petreas et al. 1991, Stephens et al.1992, 1995). Soil intake depends on animal numbers per free ranging area. Itincreases if more animals are kept per area (Schuler et al. 1997a).

37

3.4. Rabbit

The EU production of rabbits is essentially intensive. Animals are kept inwire cages, and therefore have no access to other material than the offereddiet. However, rabbits raised in wooden cages or lactating does raised incages fitted with wooden nest for their progeny can chew the surroundingwood material, which, if treated with chlorinated preservatives, couldrepresent an additional source of contamination.

Rabbit feeding is mainly based on compound feeds, which are essentiallycomposed of vegetable (plants) materials including a large proportion ofcereals, by-products (from cereals, seeds and sugar), dehydrated alfalfa meal,straw and mineral supplements. Similar ingredients are used for does andfattening rabbits. Low or high fat content are also used in feeds mainly fromvegetable origin (oil), and mineral matter can contain technological agents tofacilitate pelleting of compound feed.

There are slight variations between the diet of rabbits at thegrowing/finishing stage and the diet of reproductive animals (doe) whichtypical compositions are given.

Table 9 Typical diets for rabbits (for meat production (diet n° 1 & 2), forreproduction purposes (diet n°3) expressed in percentage

Diet composition (in percentage)Feed materials

1 2 33 3 3

A1 Roughage17 15 15

A3 Cereals 25 29 33Soybean meal 8 8 8

A4 By-products ofplant origin Other by-products 42 36 36

B2 Fat* 1 5 1C Premix** 4 4 4

* fat can include fat of animal origin and/or of vegetable origin. However for furthercalculation, it will be considered as exclusively made of animal fat.

** includes minerals, trace elements, vitamins and other feed additives

38

3.5. Fish

Fish diets are well defined and are made of fish meal, fish oil and vegetables,each of these feed materials varying depending on the target fish.

Fish meal and fish oil are essential dietary feed ingredients for all industry-produced aquafeeds for carnivorous fish, and to a lesser extent, foromnivorous fish and shrimp species. In 1992, 61.2 % of the fish meal used inaquafeeds was consumed by carnivorous fish species such as salmon, trout,eel, yellow tail and sea bream, 32% by shrimp whilst 6.8% was consumed byomnivorous species (Tacon, 1993). The highest fish meal production percountry was found in Peru (1.2 million tonnes), followed by Chile, Japan,USSR and USA. Denmark was the highest fish meal producer within the EU(260 000 tonnes), followed by Spain (135 000 tonnes) and the UnitedKingdom (51 000 tonnes). It was estimated that between 816,000 and873,000 tonnes of fish meal and between 190,000 and 205,000 tonnes of fishoil were used in aquafeeds in 1990. This is equivalent to between 13 and15% of the total world supply of fish meal and fish oil. The fish meal usagewas expected to increase by 50% to ca. 1.2 million tonnes in 2000 and fishoil usage by 77% to ca. 0.36 million tonnes per year. Assuming that globalfish meal and fish oil production will remain at their present level, thiswould mean that in 2000 about 20% of the total world supply of fish mealand fish oil would be consumed by aquafeeds. Recent data confirm theseexpectations (Eurostat, 1999). The total volume of fish meal used in Europein 1997 was 1,410,000 tonnes.

Dietary inclusion levels of fishmeal usually range from 4-5% in productiondiets for channel catfish to 75% in larval diets for marine finfish and eel orturbot production diets. Similarly, dietary fish oil supplement levels rangefrom 1-2% within production diets for omnivorous species to 15-30% withinexpanded salmon diets. Examples of a carnivorous species diet (e.g. salmon)and a non-carnivorous species diet are given in table 10. Fish meal and fishoil are not used in herbivorous fish species diets.

Table 10 Description of the usual diets for omnivorous and carnivorous fishspecies in percentage

Omnivorousspecies

Carnivorousspecies

Feed materials Fish diet expressed in %A3 Cereals 30 11

Oilseed meal 56 7A4 By-products of plant

origin Maize gluten meal - 5Fish meal 10 50B1Fish oil 2 25

C Premix* 2 2* includes minerals, trace elements, vitamins, single cell proteins and other feed additives

39

4. CONTAMINATION LEVELS OF FEED MATERIALS - DATA AVAILABLE

4.1. Detection and analysis - Technical considerations

Dioxins are normally found as complex mixtures in varying composition inthe different matrices. This requires a highly sophisticated analysis, becauseit is indispensable to separate the toxic congeners (bearing 2,3,7,8- chlorinesubstitutions) from the non-toxic. Usually, PCDDs/PCDFs are determined bycapillary-GC/MS (gas chromatography / mass spectrometry) methods.Whereas many environmental samples (as soil or sewage sludge) can beanalysed with low-resolution mass spectrometry, measurements of feed, foodand human milk or tissue samples have to be performed at ultra-trace levels.Therefore, if food or feedingstuff samples are to be analysed reliably in therange of the normal background contamination, the application of high-resolution mass spectrometry is necessary.

For comparison of analytical results, the limit of detection (lowest limit forqualitative identification, without possibility to quantify the amount) and/orlimit of determination (lowest limit for quantification) have to be taken intoaccount. Analytically, all 17 congeners with 2,3,7,8-substitution must bedetermined. For calculation of the TEQ value, the results of each of thesecongeners is multiplied by the specific TEF factor and then added up. Inmost cases, a few of the 17 congeners are below the limit of detection and/orlimit of determination. This can become critical if many congeners are notdeterminable or if the toxicological most important congeners are not found.Some laboratories are used to calculate the contribution of not detectablecongeners to the TEQ as “0”. As a consequence, low dioxin contents couldhave been the result of really low levels of the sample or of insufficientdetection/determination limits, without considering these detection/determination limits in the final TEQ calculation. To make sure that lowdioxin levels are really the result of low levels in the sample, the concept oftolerances “as upper bound limit of detection” or “upper bound limit ofdetermination” was developed. This concept demands the inclusion of thefull limit of detection or determination instead of “zero” for not detectablesubstances. It should be applied generally, with a clear preference of “upperbound limit of determination” rather than upper bound limits of detection.

A practical example should demonstrate the problem: table A16 summarisesan overview of dioxin contents in kaolinitic clay from available reports. Formontmorillonite/bentonite, a laboratory has found < 1.9ng I-TEQ/kg. Thus,obviously dioxins were not detectable with a limit of determination of 1.9ngI-TEQ/kg. For the same group, a laboratory of the same country (maybe thesame laboratory?) found 1.7ng I-TEQ/kg. It remains unclear whether therange of 1.5 to 2ng I-TEQ/kg was the practical limit of determination of thatlaboratory. This may be acceptable in a crisis situation (e.g. after the firstfinding of the contamination of kaolinitic clay) to see whether there are otherhighly contaminated samples, as well. However, these values cannot be usedfor definition of the background contamination, as the applied method isobviously not suitable for this purpose. Moreover, the method is not suitablefor determinations in the range of the tolerance of 0.5 ng WHO-TEQ/kgwhich was set in response to the finding of this contamination.

40

When the limits of determination are high for the decisive congeners, itresults in high numbers of TEQ. This has to be considered for the definitionof background contamination. Especially the use of low-resolution massspectrometers in feed materials/feedingstuffs analysis or a low weight-inquantity of a sample (for a quick and easy analysis) can cause relatively highvalues of dioxin contents as upper bound limits of determination. Thiscannot be seen from a reported TEQ level without knowledge about theresults of the individual congeners. Thus, for definition of a backgroundcontamination, published data must be reviewed critically to avoid thatrelatively high values are included only as a result of insufficient detectionlimits.

The Committee is in line with the recommendations of the report of an EUmission to Brazil (of July 1999) in connection with the citrus pulp case(European Commission, 1999c). For the control of the tolerance of 0.5ngWHO-TEQ/kg (as upper bound limit of determination) in citrus pulp andlime, it was demanded that the registered laboratories should be able to meetthe following requirements:

− demonstration of the performance of a method in the range of thetolerance, e.g. 0.5x, 1x and 2x the tolerance

− limit of detection should be in the range of about one fifth of thetolerance, to make sure that acceptable coefficients of variations are metin the range of the tolerance

For feed materials of vegetable origin in general, a limit of determination ofapproximately 0.1ng WHO-TEQ/kg dry matter seems to be appropriate todifferentiate reliably between samples with elevated dioxin levels andbackground contamination. However, to follow time trends of backgroundcontamination in those matrices, the limit of determination should be at leasta factor of 10 lower. For animal products of land origin and for fish and fishproducts, a limit of determination of 0.1ng WHO-TEQ/kg fat (as upperbound limit of determination) seems to be appropriate for both the abovementioned goals.

The qualification of the laboratories is an essential parameter. So far,laboratories applying GC/MS methods have shown that reliabledetermination of all 17 PCDDs/PCDFs congeners with 2,3,7,8-substitution ispossible even in ultra trace levels. However, the successful participation inintercalibration studies for e.g. soil or sewage samples does not necessarilyprove the competence also in the field of food or feedingstuff samples withtheir lower contamination range. Another difficulty arises from the lack ofreference material with certified dioxin content which would be helpful forinternal quality control of laboratories. So far, no such material is available.With this regard, the Committee welcomes the call published recently by theEuropean Commission (2000/C 290/05) for proposals for indirect RTDactions under the specific programme for research, technologicaldevelopment and demonstration on competitive and sustainable growth7,

7 OJ n° C 290 of 13.10.2000, p. 4

41

which invites for submission of proposals to support the development ofCertified Reference Materials, in particular in the field of environmentalcontaminants in food and animal feed (topic IV.20) and specifically forPCDDs, PCDFs and PCBs in food and feed products.

Bioassays were developed as rapid screening methods. These bioassays cangive only a TEQ range as final result, but do not give any congener specificvalues which would allow a pattern discussion. To evaluate the reliability ofmethods, numerous statistical data must be presented as the limit ofdetermination or the coefficient of variation at different relevant levels.Numbers on false-positive or false-negative results are important criteria forscreening tests, as well. Last but not least use of these bioassays incollaborative studies could demonstrate the performance. So far, there is alack of data to prove the general comparability of the bioassay results towell-documented GC/MS methods.

In the past, PCB analyses mainly focused on the determination of the markercongeners 28, 52, 101, 138, 153 and 180, which are the predominant PCBcongeners found in humans and food stuffs of animal origin. However, thetoxicity of these PCB congeners is relatively low. On the other hand, data forthose dioxin-like PCB congeners, which are much more important becauseof their toxicological potencies, are still scarce.

4.2. Evaluation of the contamination of feed materials

In order to use consistent figures for the calculations of dioxin contents ofeach animal category diet (chapter 5), low and high contamination levels aswell as a mean value of dioxin burden of the feed materials have beenestablished on the basis of the data available. These data are available inAnnex A. The figures are expressed in ng WHO-TEQ / kg dry matter. As thedatabase for dioxin-like PCBs is scarce, low, mean and high values werederived only for dioxins.

It must be underlined that the choice of the highest values was somewhatarbitrary due to the scarcity of the data and limited information on the originand conditions of analysis of the samples. This choice was made consideringvery severe in not the worst situations.

4.2.1. Vegetable feed materials

4.2.1.1. Roughage (forages, conservates and cereal straw)

Very limited data are available concerning the contamination of grass andother fodders. Table A3 summarises data (European Commission, 1999a)concerning dioxin level in some European rural, non-contaminated areas aswell as in areas known to be contaminated. The values reported range from0.13 to 2.1 ng WHO-TEQ/kg DM. Other results, obtained recently byMalisch (2000b, see table A1), show values ranging from 0.04 to 0.51 fornon-contaminated samples and from 0.84 to 24.1 for contaminated samples(clearly identified as not accidentally polluted).

42

On the basis of the available data for roughage, low level, high level andmean level retained for calculation were identified and are respectively 0.1,6.6 and 0.2 ng WHO-TEQ/kg dry matter

One paper with recent data concerning grass silage is available that indicatesmean contamination levels of 0.201 ng I-TEQ/kg DM (upperbound detectionlimit) and of 0.137 ng I-TEQ/kg DM (limit of determination), withrespective ranges of 0.120-0.650 and 0.070-0.630 (Monitoring Schleswig-Holstein, 1998), which are consistent with the preceding data for forages.Therefore, taking into account the scarcity of the information, the levelschosen for forages have been retained for the processed materials (silages,hay) as well:

Low 0.1 ng WHO-TEQ/kg dry matter

Mean 0.2 ng WHO-TEQ/kg dry matterRoughage

High 6.6 ng WHO-TEQ/kg dry matter

Remark: As sun-curing and silaging are not polluting operations per se, (seeprevious comments in chapter 3.1, introduction), soil may represent anadditional contamination source.

4.2.1.2. Roots and tubers

No data are available apart from two analyses performed on manioc reportedby Schöppe and Kube-Schwickardi (1998, see table A6), and two analysesfrom Malisch (Malisch 2000b, see table A1). As the absorption and transportof dioxins by the plants is very limited, it can be anticipated that thecontamination of roots and tubers is very low if the soil is not contaminated.Moreover the use of these materials in pig and ruminant diets is dependenton local resources and limited to certain period of the year. Therefore nolevels have been identified.

However, depending on the weather conditions prevailing at the croppingperiod, the contamination of roots and tubers by soil may be important.Namely manioc (used as starch source for energy) could be contaminateddepending on the harvesting conditions that may convey important quantitiesof soil, but also processing (drying) conditions. This is illustrated by thevalue of 0.71 ng I-TEQ/kg dry matter quoted by Schöppe and Kube-Schwickardi (1998).

4.2.1.3. Cereals, seeds, by-products of plant origin

A compilation of the data available is given for cereals and seeds (see tablesA1 and A4) and by-products of milling (table A5), starch industry (table A6),sugar industry (table A7), brewery/distillery (table A8) and oil industry (tableA9). They are mostly expressed as ng I-TEQ/kg feed. Although only fewdata are available, they show that on the average dioxin contamination of

43

these materials is relatively low (less than 0.1 ng WHO-TEQ/kg product ordry matter).

More data are required to assess the possible risk of some by-products inanimal nutrition, especially for some milling by-products, manioc, sugar beetpulp, molasses and cereal dusts.

In some cases, such as grain dust, higher values were measured (barley dust:0.330; wheat dust: 0.440 ng I-TEQ/kg DM) which confirms the prevalenceof the deposit of dioxins on externally exposed plant organs. However, theaverage of milling by-products (not specified 0.671 ng I-TEQ/kg DM; Ruoffet al. 1999) should not be over-estimated because of a contaminated sample(5.528 ng I-TEQ/kg DM).

Some references (European Commission, 1999b, Schöppe et al. 1997)indicate higher values for sugar beet pulp (0.42-0.56 ng I-TEQ/kg; see tableA7).

As a summary, the following values have been retained:

Low 0.01 ng WHO-TEQ/kg dry matter

Mean 0.1 ng WHO-TEQ/kg dry matterCereals and seeds

High 0.4 ng WHO-TEQ/kg dry matter

Low 0.02 ng WHO-TEQ/kg dry matter

Mean 0.1 ng WHO-TEQ/kg dry matterBy-products ofplant origin

High 0.7 ng WHO-TEQ/kg dry matter

4.2.1.4. Vegetable oils

Referring to the data on dioxin contents of seeds and oil by-products (cakes)(see 4.2.1.3.), i.e. feed materials derived from oleaginous plants with highlipid contents or at least residual oil (cakes), the background level is below0.1 ng WHO-TEQ/kg DM. This indicates a very limited contamination ofthese plants. Moreover refinery methods contribute to the elimination ofthese contaminating compounds. Results of Malisch (2000b) confirm thevery low contamination of vegetable oils (below 0.1 ng WHO-TEQ/kg offat). Other data are scarce.

The SCOOP task 3.2.5 provides a few analytical results from only threecountries. When compared to results of Malisch (2000b), they appear to bemuch higher. However it is indicated that these values are around the limit ofdetermination of the method used, which lead to an overestimation of thedioxin contents.

44

Finally the SCAN has retained the following figures:

Low 0.1 ng WHO-TEQ/kg dry matter

Mean 0.2 ng WHO-TEQ/kg dry matterVegetable oil

High 1.5 ng WHO-TEQ/kg dry matter

4.2.2. Feed materials of animal origin

4.2.2.1. Fish meal

The data presented in table A10 show that fish can be identified as animportant source of contamination of feedingstuffs by dioxins, PCBs anddioxin-like PCBs.

Few dioxin analyses have been carried out in fish meal and fish oil. Possiblymore have been carried out within the industries but have not beenpublished. The above mentioned table gives an overview of dioxinconcentrations found in fish meal and fish oil samples.

The data are relatively consistent. There is a clear difference incontamination levels between fish meal originating from the (south) Pacific(Chile and Peru) and those originating from European waters. Theconcentrations measured in the latter are 8 fold higher than in the former.

Within Europe no clear difference could be found between fish mealoriginating from northern waters, e.g. around Iceland and the North Sea. Theavailable database may be too small to see such differences. In some samplesthe dioxin levels have been expressed as Nordic-TEQ or I-TEQ, whichmeans that the TEF values used for dioxins and furans are slightly differentthan those used in the WHO-TEQ calculation, and, more important, PCBlevels have not been taken into account. The data of Vartiainen andHallikainen (1995) show an average factor of 4.0 between PCB and dioxinTEQ values in fish oil. A similar ratio was found by de Boer et al. (1993) inDutch fish samples. Malisch (1996) reported PCB TEQ/dioxin TEQ ratios of2.7-23.3 in freshwater fish (eel, pike and roach), with mean ratios of 6-8.However, some eel may have contained exceptionally high PCB levels.

The following ranges and means expressed as dioxin in ng WHO-TEQ/kghave been retained for fish meal (including only PCDDs/PCDFs):

South Pacific (Chile, Peru) Europe

Range 0.18 - 2.1 ng/kg fat 0.3 - 47 ng/kg fatMean 1.17 ng/kg fat 10.1 ng/kg fat

45

Considering that the fat represents approximately 12% of the fish meal, thefollowing values (for PCDDs/PCDFs only) expressed in ng WHO-TEQ / kgdry matter (DM) are retained:

Fish meal originating from the South Pacific area (Chile, Peru):

Low 0.02 ng WHO-TEQ/kg dry matter

Mean 0.14 ng WHO-TEQ/kg dry matterFish meal

High 0.25 ng WHO-TEQ/kg dry matter

Fish meal originating from the European area:

Low 0.04 ng WHO-TEQ/kg dry matter

Mean 1.2 ng WHO-TEQ/kg dry matterFish meal

High 5.6 ng WHO-TEQ/kg dry matter

To take into account the contribution of PCBs to the total TEQ, the TEQresults reported by references Döring (2000), Anon (2000), Lundebye-Haldorsen and Lie (1999), Guodmundsson (1999), Andunsson (2000), andAnon (1999a) in the table A10 have been multiplied by a factor of 5.Original and corrected data are presented in that table.

The following ranges and means expressed as dioxin + PCB WHO-TEQ/kgon a dry matter (DM) basis have been calculated for fish meal:

Fish meal originating from the South Pacific (Chile, Peru) area:

Low 0.11 ng WHO-TEQ/kg dry matter

Mean 0.7 ng WHO-TEQ/kg dry matterFish meal

High 1.26 ng WHO-TEQ/kg dry matter

Fish meal originating from the European area:

Low 0.18 ng WHO-TEQ/kg dry matter

Mean 6.1 ng WHO-TEQ/kg dry matterFish meal

High 28.2 ng WHO-TEQ/kg dry matter

46

4.2.2.2. Fish oil

The picture for fish oil is similar to that of fish meal: higher contaminationlevels in European fish oil compared to fish oil of South Pacific (Chile, Peru)origin. By using the data from the table A14, the following ranges for dioxinhave been retained (expressed in ng WHO-TEQ/kg fat):

Fish oil originating from the Pacific area:

Low 0.16 ng WHO-TEQ/kg fat

Mean 0.61 ng WHO-TEQ/kg fatFish oil

High 2.6 ng WHO-TEQ/kg fat

Fish oil originating from the European area:

Low 0.7 ng WHO-TEQ/kg fat

Mean 4.8 ng WHO-TEQ/kg fatFish oil

High 20 ng WHO-TEQ/kg fat

The same correction has been applied for the PCB concentrations in somefish oils (Anon (2000), Guodmunsson (1999), Andunsson (2000) and Anon(1999a)).

Therefore, the following ranges for dioxin + PCB TEQ data in fish oil(weighted means by using the data from the table A14) expressed in ngWHO-TEQ per kg fat have been calculated:

Fish oil originating from the Pacific area:

Low 0.8 ng WHO-TEQ/kg fat

Mean 3 ng WHO-TEQ/kg fatFish oil

High 13 ng WHO-TEQ/kg fat

Fish oil originating from the European area:

Low 3.5 ng WHO-TEQ/kg fat

Mean 24 ng WHO-TEQ/kg fatFish oil

High 100 ng WHO-TEQ/kg fat

47

4.2.2.3. Meat and bone meal, animal fat

As dioxins are bioaccumulated in animal fatty tissues, one could anticipatethat recycling of animal products through rendering would give an efficientconcentration process. The analytical results concerning animal fats (seetable A13) indicate a range of values (0.5 to 1.5 ng WHO-TEQ/kg of fat)that are higher than those found in most other feed materials of plant origin,but still much lower than in fish products.

The data concerning meat and bone meals, which contain only 7 to 8% of fat,are summarised in table A11. They indicate a low contamination withdioxins. Feathers collected separately at bird slaughter constitute, onceprocessed, a protein source and are used as feed. The dioxin contaminationlevels of feather meals are similar to those of meat and bone meal.

On the basis of these data the levels of contamination retained by the SCANare:

Low 0.5 ng WHO-TEQ/kg fat

Mean 1 ng WHO-TEQ/kg fatAnimal fat

High 3.3 ng WHO-TEQ/kg fat

Low 0.1 ng WHO-TEQ/kg dry matter

Mean 0.2 ng WHO-TEQ/kg dry matterMeat and bone meal

High 0.5 ng WHO-TEQ/kg dry matter

4.2.2.4. Milk replacers

Milk constitutes an efficient elimination pathway of dioxins in animals,which is reflected by the average contamination found in milk products.However information on the data presented in table A12 are not sufficient toestablish levels of contamination.

Dairy products, such as dried whey and skimmed milk, used in milkreplacers are defatted materials and therefore considered not contaminated.However, milk replacers are generally based on refatted dairy products(about 20% of fat on dry matter basis) such as refatted whey or refattedskimmed milk. Mixtures (50/50) of animal fat and vegetable oil are used.Therefore, considering that fat represents ca 20% of the milk replacers, thefollowing values are proposed by SCAN for calculation.

Low 0.06 ng WHO-TEQ/kg dry matter

Mean 0.12 ng WHO-TEQ/kg dry matterMilk replacers

High 0.48 ng WHO-TEQ/kg dry matter

48

4.2.3. Soil and other feed materials

4.2.3.1. Soil

Many studies have been aimed at identifying “hotspots” of contamination,and as a result such locations have been more intensively sampled andanalysed than background or base line locations. Data obtained fromdifferent European countries have been compiled (European Commission,1999a). As expected a wide range of concentration values for dioxins wasfound. It has not been possible to carry out any statistical analysis ofavailable data, as countries or individual reports provided aggregated datacovering varying numbers of samples, sampling procedures, periods andlocations. However, the analysis of the data shows that in most cases it isimpossible to distinguish significant differences in backgroundconcentrations of dioxin in rural and urban locations. Major differences wereonly noted between contaminated and not contaminated sites. Table A15summarises the range of dioxin concentrations measured in pasture andarable soils in different European countries.

Taking into account the grazing and rearing practices, the potentialcontamination of the animal feed resources is generally below 3 to 4 ng I-TEQ/kg DM, except for the contaminated areas where a factor of x 1,000may be encountered. Facing the difficulty to choose a typical concentrationvalue for a "naturally" but highly contaminated soil, the highestconcentration appearing in table A15 has been retained as a conservativeapproach.

Low 0.5 ng WHO-TEQ/kg dry matter

Mean 5 ng WHO-TEQ/kg dry matterSoil

High 87 ng WHO-TEQ/kg dry matter

4.2.3.2. Binders and anticaking agents

The available data are shown in table A16. The dioxin contamination ofthese various minerals is very low. Referring to chapters 2.1 and 2.3, it isclear that the contamination is highly dependent on the geological conditionsprevailing at the mining site, and that the accidents which occurred withkaolinitic clay, ball clay or bentonite are the exception. The upper levelretained for this category of feed materials is the maximum permitted levelaccording to Commission Regulation n°2439/99 of 17 November 19998 i.e.0.5 ng WHO-TEQ /kg dry matter. The other values have been chosenaccording to the data listed in table A16.

Low 0.1 ng WHO-TEQ/kg dry matter

Mean 0.2 ng WHO-TEQ/kg dry matterBinders andanticaking agents

High 0.5 ng WHO-TEQ/kg dry matter** not analysed

8 JO n° L 297 of 18.11.1999, p. 8

49

4.2.3.3. Trace elements / Macrominerals / Premixes

The very limited available data are presented in table A17. As premixes are adilution of trace elements/macrominerals and other ingredients with feedmaterials like cereals or by-products used as a carrier, their dioxin content iscomparable for both categories. Therefore similar values have been retained:

Low 0.1 ng WHO-TEQ/kg dry matter

Mean 0.2 ng WHO-TEQ/kg dry matterTrace elements,macrominerals

High 0.5 ng WHO-TEQ/kg dry matter** not analysed

Low 0.02 ng WHO-TEQ/kg dry matter

Mean 0.2 ng WHO-TEQ/kg dry matterPremixes

High 0.5 ng WHO-TEQ/kg dry matter** not analysed

4.2.4. Evaluation of the levels of contamination

On the basis of the available data, and the calculated mean value, it can beconcluded that the most important contaminated feed materials used inanimal feedingstuffs are, by order of importance:

(1) Fish oil and fish meal

(2) Animal fat

(3) All other feed materials.

Soil, although not being stricto sensu a feed material, is being ingested bysome categories of animals. The soil contamination is very variable, butcould become a significant contributor to animal exposure to dioxins, as itscontamination can be high.

4.3. Estimation of background contamination of feed materials

All data summarised in table A3 to A17 have been used to identify the levelof contamination of the different feed materials and show a wide range ofdioxin contents. However, most data (with the exception of the data providedby Malisch, 2000b, tables A1 and A2) do not contain information on twopoints important for the interpretation of background levels:

− the upper bound limit of determination of the method (see chapter 4.1.)

− whether a possibly contaminated feed material is included

Thus, it is difficult to conclude the range of the normal backgroundcontamination from tables summarising all published results and to separatethis from elevated levels.

50

4.3.1. Estimation of background contamination on the basis of feedmaterials data

To give an orientation in the discussion of the background contamination offeed materials, the Committee chose therefore to use the detailed data of theMalisch study (Malisch, 2000b; Malisch and Fürst, 2000). Here the results ofdioxin determinations in 245 feed materials samples are evaluated. Thesewere analysed between 1993 and 1999 in Germany. Almost 80 % of thesesamples were analysed in 1998 and 1999 in relation to episodes of severedioxin contamination of feedingstuffs.

From all 245 analysed samples, whose origin was fully identified, 72samples were related to a specific dioxin contamination (Brazilian citruspulp pellets; kaolinitic clay; hay from PCP-treated storage rooms; grassfertilised with contaminated Thomas phosphate). The remaining 173 sampleswere considered to reflect the background contamination and evaluatedstatistically. When compared to the number of feed materials and ingredientsthat can enter into the composition of feedingstuffs, only a very limitednumber of analyses per compound are available. Therefore the collectedsamples were grouped in three categories and a limited of sub-groups asexplained in chapter 3. In addition to the three categories (A, B and C), dataare given concerning dioxin concentrations in commercial compound(mixed) feeds (feedingstuffs: category D).

The data of the analysis of the different feed materials groups, separated forcontaminated and not contaminated samples and recalculated for the WHO-TEQ content as upper bound determination limit (with inclusion only ofPCDDs/PCDFs), are presented in table A1.

The frequency of the distribution of dioxin contamination of feed materialsgroups A1 to A4 is shown in Figure 1.

It is obvious that the usual background contamination of products of thegroups A2, A3 and A4 is below 0.2 ng WHO-TEQ/kg dry matter (air driedor at 105 °C dried; upper bound determination level; only PCDDs/PCDFsincluded).

Feed materials of the group A1 (forages, conservates and cereal straw) havea tendency to slightly higher levels. This is the result of numerous grasssamples which are here included. Grass and leaves provide a large surfacearea for aerial contamination with dioxins on its wax layer, which is able toabsorb PCDDs/PCDFs. In comparison to grass, feed materials of the groupsA2 and A3 have a smaller surface, with tubers and roots being contaminatedpredominantly by soil and cereals and seeds by aerial deposition. The timefor growth (season of the year) could have an influence on the dioxincontent, as the growth rate is different. To take into account as manyparameters as possible, the grass, hay and grass silage samples were taken atdifferent times and reflect rural and highly populated areas, however withoutknown dioxin emittents in the surrounding area. Thus, they do not includehighly industrialised areas e.g. with steel production.

51

Figure 1

0

2

4

6

8

10

12

14

16

18

no. o

f sam

ples

< 0,05 < 0,10 < 0,15 < 0,20 < 0,25 < 0,30 < 0,35 < 0,40 < 0,45 < 0,50 < 0,55 < 0,60 < 0,65

A2A3

A4A1

ng WHO-TEQ/kg (air-)dried matter(upper bound determination limit)

Dioxins in feed materialsA1: forages and conservatesA2: tubers and rootsA3: cereals and seedsA4: by-products from plant origin

Malisch, R. CVUA Freiburg, 2000

52

Figure 2 shows that most compound feeds are below 0.1 ng WHO-TEQ/kg(air-) dried matter, with a limited numbers of samples having a dioxincontent up to 0.25 ng WHO-TEQ/kg.

Figure 2

As a result, for feed materials of groups A2 to A4 or compound feeds a usualbackground contamination is in the range <0.1 to <0.3 ng WHO-TEQ/kg(air-dried) matter. Grass, hay and grass silage have a tendency to slightlyhigher dioxin levels. Almost all samples from the group A1 from rural andhighly populated areas are below 0.5 ng WHO-TEQ/kg DM, however, moredata from other areas in Europe would be necessary to support the overview.

Figure 3 summarises the results of mineral feeds. Also here, the majority ofsamples have a dioxin content below 0.1 ng WHO-TEQ/kg. Most sampleswere analysed shortly after the first reports about the newly detected dioxinsource in highly contaminated kaolinitic clay. At that time, somemanufacturers of mineral feeds had different stocks of raw material. Mineralfeeds from the market from the same producer could have a wide range ofcontamination. Thus, samples below 0.5 ng WHO-TEQ/kg were considered“not contaminated”. However, it can be anticipated that after exclusion ofthe contaminated products, the frequency distribution in the range between0.1 and 0.5 ng WHO-TEQ/kg would shift to lower dioxin levels.

0

2

4

6

8

10

12

14

16

no. o

f sam

ples

< 0,05 < 0,10 < 0,15 < 0,20 < 0,25 < 0,30 < 0,35 < 0,40 < 0,45 < 0,50 < 0,55 < 0,60 < 0,65

S1

ng W HO-TEQ/kg (air-)dried m atter(upper bound determ ination lim it)

Dioxins in feed ingstuffD: com pound feed

M alisch, R. CVUA Freiburg, 2000

53

Figure 3

01234567

no. o

f sam

ples

< 0,05 < 0,10 < 0,15 < 0,20 < 0,25 < 0,30 < 0,35 < 0,40 < 0,45 < 0,50 < 0,55 < 0,60 < 0,65

C

ng WHO-TEQ/kg (upper bound determination limit)

Dioxins in feed materialsC: minerals

Malisch, R. CVUA Freiburg, 2000

The feed materials of animal origin including 19 fat samples and one milksubstitute are shown in figure 4. Besides one sample, feed materials of animalorigin had a dioxin contamination below 2ng WHO-TEQ/kg fat (upper bounddetermination level; only PCDDs/PCDFs included).

Figure 4

02468

10

no. o

f sam

ples

< 0.5 < 1.0 < 1.5 < 2.0 < 2.5 < 3.0 < 3.5 < 4.0 < 4.5 < 5.0 < 5.5 < 6.0 < 6.5 < 7.0B

ng WHO-TEQ/kg fat (upper bound determination limit)

Dioxins in feed materialsB: feed materials of animal origin

Malisch, R. CVUA Freiburg, 2000

Most samples were below 1 ng WHO-TEQ/kg fat. One sample had a dioxincontent of 6.64 ng WHO-TEQ/kg fat. The composition of the samples is notknown as feed materials were sampled randomly at certain occasions,

54

without having the composition of these. It cannot be excluded that thesample with a high dioxin content contained fish oil. For a broad overview, itis therefore necessary to include the contribution of fish oils, as well.

4.3.2. Estimation of background contamination of feed materials ofvegetable origin, on the basis of food data

Considering that vegetables used as food and plants cultivated as feedmaterials are exposed to the same dioxin contamination conditions, theanalysis of data available for food vegetables can be used to drawconclusions concerning feed materials of plant origin. Regarding thedifferent contamination pathways, the differentiation of vegetables into thefollowing four groups is useful (Malisch, 1998):

- Group 1: edible part growing underground reflecting the soilcontamination (potatoes, carrots, onions, celeriac, beetroot)

- Group 2: edible part (fruit, tubers or others) growing on the groundreflecting basically the soil contamination (e.g. strawberries,zucchini (courgette), kohlrabi, celery, cauliflower, rhubarb)

- Group 3: leafy vegetables growing near to the ground reflecting thecontamination of soil and air (e.g. lettuce, sugarloaf, endive,savoy cabage, leek, white cabbage, kale)

- Group 4: fruit growing distant to the ground reflecting the aircontamination (e.g. apples, tomatoes, aubergines, paprika)

Table A2 summarises the updated results of all food samples reflectingbackground contamination. They were recalculated as ng WHO-TEQ/kg drymatter (upper bound determination limit; only PCDDs/PCDFs included)(Malisch, 2000b).

It can be concluded that:

• Generally, the background contamination of vegetables is in the range <0.1 to < 0.3 ng WHO-TEQ/kg DM (upper bound determination level;PCDDs/PCDFs only).

• The only vegetable with a considerable soil-plant transfer ofPCDDs/PCDFs is zucchini (Hülster, 1994; Hülster et al, 1994, Neumannet al, 1999). This is also obvious from the results of group 2. Thisobservation stresses the fact that physiological specificities exist and thatunexpected contamination cannot be excluded a priori from othervegetable materials not considered. In relation to its weight, kale has avery big surface with wax layers, which can absorb dioxins from the airmuch more efficiently than vegetable with a smaller surface. Thus, kale isa special kind of food, which illustrates the air-plant transfer.

• The dioxin content of potatoes for food use is in the range <0.1ng WHO-TEQ/kg dry matter. As there is a lack of data for potatoes by productsused as feed materials, the dioxin content of these can be assumed to besimilar to the dioxin content of potatoes for food use.

55

4.3.3. Estimation of background contamination of feed materials of animalorigin, on the basis of food data

The following products of animal origin are used as feed materials:

• Fish oil, fish meal

• Animal fat, meat and bone meal

• Milk products

It can be assumed that the dioxin content in feed materials of animal origin isnot different from the dioxin content of products of animal origin intendedfor human consumption. As there are many data available from the officialfood control, these data can be included to widen the overview. Thecompilation of EU dioxin exposure and Health data summarises theconcentrations of dioxins in foods. For each sort of food, a wide range ofdioxin concentrations is reported (European Commission, 1999a). Thus, thesame questions as asked in the above chapter 4.1 must be answered forsetting the background contamination in food of animal origin.

Additionally, it must be taken into account that the dioxin content in foodand human samples has decreased considerably in the past 10 years. Thus,data from the beginning of the 90s cannot be used for setting the backgroundcontamination at the beginning of the new millennium.

The available information on the occurrence of dioxins in food of animalorigin (animal fat, meat, milk (products)) is in the order of 1 ng I-TEQ/kg fat(SCOOP Task 3.2.5)

The dioxin content in fish is generally considerably higher. However, thelevels vary considerably for two reasons. First, the dioxin contamination isdifferent in different areas. Second, the level is dependent on the fat contentof the fish which varies extremely between species (between 0.04 % for apike and 40 % for an eel) and also between seasons. Because of theaccumulation of PCDDs/PCDFs in fatty tissue, the extreme different fatamounts can cause extremely different dioxin levels when correlated to freshweight or fat base.

56

5. TOTAL CONTAMINATION ESTIMATES OF TYPICAL DIETS AND IDENTIFICATION OFTHE MAIN FEED MATERIALS CONTRIBUTING TO THE CONTAMINATION.

Table 11 summarises the dioxin contents of the main feed materials established bythe SCAN on the basis of the available data submitted by Member States to theEuropean Commission or published, or referring to maximum permitted levelsaccording to current European legislation.

It includes the identified “low” and “high” levels and the mean level fixed by theSCAN, as basis to estimate the total dioxin content of each specific species diet. Itmust be underlined that the choice of the highest values was somewhat arbitrary dueto the scarcity of data and limited information on the origin and conditions ofanalysis of the samples. This choice, made considering very severe if not the worstsituations, aimed not to give realistic estimates but to evidence which feed materialswere the most influential on the contamination of different types of diets.

Table 11 Dioxin contents of the basic feed materials evaluated by the SCAN fromthe available data (in ng WHO-TEQ/kg dry matter; with inclusion onlyof PCDDs/PCDFs).

Dioxin levels in feed materials(ng WHO-TEQ/kg DM)

Feed materials Low Mean HighRoughages 0.1 0.2 6.6Cereals and seeds (Legumes) 0.01 0.1 0.4By-products of plant origin 0.02 0.1 0.7Vegetable oil 0.1 0.2 1.5Fish meal Pacific (Chile, Peru) 0.02 0.14 0.25Fish meal Europe 0.04 1.2 5.6Fish oil Pacific (Chile, Peru) 0.16 0.61 2.6Fish oil Europe 0.7 4.8 20Animal fat 0.5 1 3.3Meat and bone meal 0.1 0.2 0.5Milk replacers 0.06 0.12 0.48Soil 0.5 5 87Binders and anticaking agents 0.1 0.2 0.5#

Trace elements, macrominerals 0.1 0.2 0.5#

Premixes 0.02 0.2 0.5#

#: according to Commission Regulation n°2439/99 of 17 November 19999

On the basis of the dioxin levels in feed materials taken from the available data andusing the percentage of the different feed ingredients in the diets, totalcontamination estimates of typical diets have been calculated. They are summarisedin Annex B. Each estimation in Annex B consists of two parts: the upper tablepresents the results of the contamination of all individual feed materials and of thetotal diet in ng WHO-TEQ/kg dry matter, the lower table the percentage of thecontribution of each material to the dioxin contamination of the diet. For all diets,the derived values for low, mean and high contamination were applied. In addition,if fish meal or fish oil is used in the diet, the calculations are done separately

9 JO n° L 297 of 18.11.1999, p. 8

57

assuming that fish meal/oil comes either from European waters or from Pacificwaters.

It should be noted that the composition of the diets is expressed as percentage of aningredient in the diet on a product basis and not on a dry matter basis, except forruminants as in this case diet includes feed materials with a high water content (suchas grass). For the other species, the different materials used are dry (e.g. cereals, fator meal) and their dry matter content is almost equivalent to the product itself (theirwater content being ca 10%). Therefore this difference between product and drymatter was not considered significant and calculations were made using the dietcomposition data on a dry matter basis for ruminants and on a product basis for theother species. The total contamination levels are however expressed in ng WHO-TEQ / kg dry matter diet.

When available, the concentration of the dioxins directly measured in complete dietswill be compared.

As the database for dioxin-like PCBs is scarce, low, mean and high values wereestablished for dioxins only. The limited data available for different food materialsof plant origin (which can be extrapolated to the corresponding feed materials) showthat the WHO-TEQ values derived from the determination of dioxin-like PCBs areroughly in the same range as those determined for PCDDs/PCDFs. However, in fishand fish products, the contribution of dioxin-like PCBs might be 5 times higher thanthat of PCDDs/PCDFs. As a result, the relative contribution of the individual feedmaterials would change significantly. The WHO-TEQ values includingPCDDs/PCDFs and dioxin-like PCBs would increase about 5 fold for fish and fishproducts, whereas for all other feed materials, the content would roughly bedoubled.

5.1. Ruminants

All calculations on ruminants are based on the use of one concentrate, amongthe four identified in chapter 3 (table 2). Using a conservative approach, theCommittee chose to use the levels of contamination identified for theConcentrate IV as it included fat of animal origin as well as fish meal, for thecalculation of the contamination of the different ruminants diets, althoughmore appropriate concentrates may be used in practice depending on the typeof production and on the feed materials available. The calculated averagecontamination of this concentrate was the highest, if fishmeal from Europe isused, although in the same range as the others. The calculation ofconcentrates contamination is presented in Annex B, tables B1 and B2.

Calculations of the contribution of the individual feed materials to thedifferent diets are presented in Annex B, tables B3 to B6 for dairy cattle, andB7 to B12 to beef cattle, which includes the particular case of veal calves(tables B11 and B12). Generally for all the different ruminant diets, the meancontamination is considered to reflect the usual background in the range ofabout 0.2ng WHO-TEQ/kg dry matter.

The relative dioxin contribution reflects the composition of the diet, whichincludes mainly roughage and concentrate with numerous variants in theirrespective rate. As roughage and concentrate IV are derived with 0.2

58

respectively 0.195 (which can even be rounded to 0.2) ng WHO-TEQ/kg drymatter, there is no effect on the dioxin content of the diet with variation ofthe composition. Changes of the composition of the concentrate show theinfluence of fish meal and animal fat: 5 % fish meal of European watersresult in 31 % of the average dioxin content of concentrate IV, and 4 %animal fat in 21 %. Thus, the use of concentrates without these feedmaterials would lower the dioxin content of the diets with high portions ofconcentrate. However, the most important factor especially is roughage, asthe comparison to low and high contaminated feed materials demonstrates.This is due to the level identified by the Committee in table 11 and leads toan increase of the diet contamination calculated in ng WHO-TEQ/kg drymatter by a factor of 10 to 30, depending on the rate of roughage in the dietcomposition. For example, in ruminant diet n°8 in table B9, the dioxincontent increases from 0.2 ng WHO-TEQ/kg dm to 4.88 ng WHO-TEQ/kgdm, and 70% of the diet ingredients (roughage) contributes to 95% of thecontamination.

In the case of veal calves, for the low and mean level calculation (see tablesB11 & B12), the diet contamination results predominantly (84%) from milkreplacer, constituting 90% of the daily ration. However the dioxin burden ofmilk replacer results from its lipid source and content. For the high levelcalculation, 10% roughage in the daily ration contributes to greater extent tothe total dioxin intake than the milk replacer.

The presence of soil in the ingested components of the diet as well as directconsumption by animals on pasture can increase the total dioxin exposure ofruminants. However, it must be noted that the bioavailability of dioxinsadsorbed on soil is lower than that of other dioxin sources. Of course,animals kept indoors, during winter for example, and more generally animalsfed with diets including higher percentage of concentrates, have a lower riskof presence of soil and dioxin originating thereof. This applies in particularto the specific production of calves and "baby beef".

5.2. Pigs

Calculations on the diet contamination of piglets are presented in Annex B,tables B13 and B14, of pigs for fattening in tables B15 and B16 and of sowsin tables B17 to B20.

Calculated on the basis of mean values of table 11, the level ofcontamination of all pig feeds is low, ranging from about 0.10 up to 0.23 ngWHO-TEQ/kg dry matter.

The derived maximum dioxin content for the average contamination (0.23 ngWHO-TEQ/kg dm) refers to piglets diets (see table B13). This is mainly dueto the use of fish meal and animal fat in the piglets' diets. Diet n°2 as anexample demonstrates the correlation. The addition of 8 % animal fat to acompound feed would result in a contribution to the dioxin content of 35 %.As well, the use of only 5 % of fish meal from European waters results in adioxin contribution of 26 % of the whole diet, whereas the same amount of

59

fishmeal from Pacific waters would give only 4 % of the total dioxincontribution. This is also seen for the other pigs' categories.

In the particular case of pig diet n°14, which includes 32% roughage and isintended for pregnant sows, the use of highly contaminated forage (instead ofmean contaminated one) can cause up to ca 80% of the dioxin contaminationof this fibrous diet. This can be explained by the figure retained by SCANfor the high level of contamination of roughage which is significantly higherthan the mean level (respectively 6.6 and 0.2) and impacts consequently onthe total contamination of this particular diet, causing a level for themaximum contamination of 2.5 ng WHO-TEQ/kg dry matter. In contrast tothis, in other cases the high contamination does not exceed 0.8 g WHO-TEQ/kg dry matter. More generally, as to be expected, the use of highlycontaminated materials multiplies the level of contamination of the diet by afactor of 4 to 5. In the case of pig diet n°14, the factor is even higher, asexplained in the above paragraph.

The consumption of significant amounts of soil for outdoor bred animals canincrease the exposure to dioxin, in particular for reproductive sows oftenbred outdoor. Although these animals are generally not used to producemeat, their progeny can be contaminated through suckling their mother untilweaning.

Wild pig diets, although not described in this report, are characterised by theconsumption of high amounts of soil, forage and roots. This may lead tohigher dioxin intake in comparison with those of domestic pigs.

5.3. Poultry

Calculations on the poultry diets' contamination are presented in Annex B,tables B21 to B26.

The dioxin concentrations of poultry diets with low contaminated feedmaterials vary between 0.022 and 0.063 ng WHO-TEQ per kg. Underconsideration of mean values of contamination (see table 11) dioxinconcentrations of feedingstuffs vary between 0.114 and 0.194 ng WHO-TEQper kg.

All diets contain less than 5 % of fish meal and up to 10 % of fat, with a sumof both ingredients being less than 10%, but the use of these two feedmaterials (5 % fish meal from Europe and 3 % of animal fat) contributes upto ca 50% of dioxin load on average (table B22, diet n° 1). This is alsoillustrated by table B26 with the diet n°7 for turkeys where fish meal andanimal fat represent only 5% of the diet composition, but contribute to 36 %of the dioxin contamination.

As a result, higher proportions of fish meal and animal fat and highercontaminated feeds (see table 11) may significantly increase dioxinconcentration of poultry feedingstuffs.

The evaluation of the mean dioxin content of poultry diets is in agreementwith data from Malisch (2000b) who analysed 14 commercial poultry

60

feedingstuffs from Germany and measured dioxin concentrations between0.012 and 0.232 ng WHO TEQ/kg.

In addition to the exposure of animals to dioxins through their diets, in thecase of free ranging animals, the soil intake may represent an additionalsource of contamination. Indeed, if between 2 and 10 g soil per layer and perday (see chapter 3.3.) is taken up, the additional dioxin intake varies between0.001 and 0.87 ng WHO-TEQ per layer and per day considering the lowestand highest dioxin concentration of soil identified in table 11.

Assuming a daily feed intake of 120 g per hen and using mean values for thecalculation of diet n°5 in table B23, the dioxin intake per day and per animaloriginating from the feed would be around 0.014 ng WHO-TEQ. Using alsothe mean contamination level of soil and a mean soil intake of 6g per day,the dioxin intake originating from soil would be 0.03 ng WHO-TEQ. So, forfree ranging layers, on the basis of the sum of dioxin contribution from dietand soil, the soil could represent a significant part (around 70%) of thedioxin exposure of animals. Bioavailability is however lower for soil andshould be considered for a proper assessment of its final impact on thecontamination of food of poultry origin.

5.4. Rabbits

Calculations on the diet contamination of rabbits are presented in Annex B,tables B27 and B28. The dioxin contamination of rabbit feeds is between0.13 and 0.17 ng WHO-TEQ/ kg dry matter and can be considered low.

There are no significant differences in the total contamination of the differentkinds of feedingstuffs used for the two types of production (meat productionand reproductive animals). This is explained by the relatively similar basiccomposition of the diets.

Table B28 shows that the contribution of the different feed materials to theoverall contamination is consistent with their respective rate in the dietcomposition. However, the addition of 5% fat of animal origin cancontribute to 30% of the diet contamination (diet n°2).

When considering high levels of contamination, roughage becomespreponderant as 20% inclusion in the diet brings ca 70% of thecontamination. This is linked, as in the case of pig diet n°14 (see chapter5.2), to the high level of roughage contamination retained by the Committee.It indicates that roughage, when highly contaminated, needs carefulconsideration.

Finally, apart from exposure to dioxins through their "classical" diet, rabbitsraised in wooden cages or lactating does raised in cages fitted with woodennest for their progeny can chew the surrounding wood material which, iftreated with chlorinated preservative, can represent an additional source ofcontamination.

61

5.5. Fish

Calculations on the fish diet contamination are presented in Annex B, tablesB29 and B30.

Clearly the dioxin contamination of the diet of farmed fish is relatively high,particularly for carnivorous species as their intake of fish meal and fish oil isquantitatively important. The origin of the fish meal and fish oil is also ofgreat influence. When using European fish oil and fish meal, the dietcontamination is ca 8 fold higher when compared to those based on fish mealand oil originating from the Pacific area (Chile, Peru). For omnivorousspecies, the differences are less dramatic because the percentage of fish mealand oil in their diet is considerably lower than for carnivorous species. Thiscan be further applied to herbivorous fish species.

The use of fish oil of various origins is dependent on various factors such asimport and export figures and market prices. In 1997 the net amount ofEuropean fish meal was ca. 30% against 70% of Pacific fish meal on a totalusage of 1.4 million tonnes. However, due to El Niño, in 1998 the amount ofEuropean fish meal used has been much higher. Clearly these changes in usehave a considerable impact on the TEQ levels in fish meal and fish oil. Thecontribution of other feed materials is only of some significance whencombined with weakly contaminated South Pacific fish meal or fish oil(using low values).

In most fish oils and fish meals, dioxin-like PCBs determine ca 80% of thetotal TEQ, although variation (from 20% to 95%) may occur. This meansthat, when PCB TEQ are taken into account in addition to the dioxin TEQ,the mean total contamination level of an average product calculated in tableB29 may increase from 1.8 to 9 ng WHO-TEQ/kg dry matter for feedmaterials (of fish origin) of European origin in the case of carnivorousspecies (for a high contaminated product, from 7.9 to 40 ng WHO-TEQ/kgdry matter). This is a considerable difference with the diet contaminationcalculated on the basis of dioxins only.

As shown in table B29, use of fish meal and oil with known low levels ofcontamination, like for instance some fish products originating from certainareas of the Pacific (Chile, Peru), can lower the global contamination.Although fish meal and fish oil still impact the most on the total dietcontamination in comparison to other feed materials (see table B30), the dietcontamination is divided by a factor of 8 (case of carnivorous species).

If the use of less contaminated feed materials of fish origin has a positiveimpact on the whole diet contamination, it is rather limited; leading to adecrease from 98% to ca 89% of the dioxin contribution in the case ofcarnivorous species. Alternative feed materials to replace fish meal and oil infish diet could contribute to lower this impact, provided these alternatives areless contaminated. There is indeed a tendency to reduce the use of fish mealand fish oil in farmed fish diets through the utilisation of plant seed oils suchas soybean meal and, to a lesser extent single cell proteins. However onemust be aware that nutritional constraints exist that limit the exercise.Finally, in order to overcome the contamination of fish oil known to be

62

heavily contaminated (case of fish oil of European origin), use ofpurification/decontamination techniques, like for example distillationmethods, are currently envisaged.

5.6. Identification of the main contributors to diets contamination

Considering the best situation, i.e. calculating diet burden using the lowlevels established by the SCAN for all feed materials, it appears that,whatever animal species is considered, all different diets exhibit lowcontamination levels below 0.1 ng/kg diet, except for carnivorous fish fedfish oil and fish meal from European origin for which this value is around0.2 ng/kg diet.

On the basis of the mean values retained by the SCAN that reflect theaverage situation, whatever animal species is considered, the whole dioxinburden is comprised between 0.1 and 0.3 ng/kg diet, except for carnivorousfish that receive fish meal and fish oil of European origin. In that case dioxincontamination reaches 1.8 ng/kg diet.

Using the high and somewhat arbitrary and overestimating dioxinconcentrations retained (see above, chapter 5.), the whole contamination isover 0.5 ng/kg diet, the ruminants and fish being the most exposed with adioxin burden far over 1 ng/kg diet. One exception is that of the omnivorousand carnivorous fish receiving fish oil and fish meal originating from theSouth Pacific area where the value remains below 0.9 ng/kg diet.

Considering the rate of inclusion of feed materials combined with theirimpact (in percentage) on the total contamination of the diets, feed materialsof fish origin (of European origin) and animal fat represent the majorcontributors to diets contamination. For instance, low quantities (5% to 8%)of these ingredients lead to around a third of the diet contamination (in pigsor poultry). When these feed materials are used together, their contributioncan represent more than 50% of the contamination.

In the case of diets made of one or two predominant feed materials (like forruminants or for fishes), the diet contamination is mainly the result of theproportion of these materials in the diet. For this reason, roughage and feedmaterials of fish origin are the respective main contributors to diets ofruminants and fishes.

As farmed fish combines an important consumption of feed materials of fishorigin (up to 75% in the diet of carnivorous species) with a high level ofcontamination of these feed materials, it is the food producing animal mostexposed to dioxins.

A special mention should be made on the contribution of soil as a significantpart of the diet of certain animals, i.e. grazing ruminants and free-rangeanimals. However its importance in terms of dioxin availability for theanimals cannot be evaluated on the basis of the present knowledge.

63

6. CARRY OVER OF DIOXINS FROM FEED TO FOOD

6.1. General findings

Generally, PCDDs/PCDFs are chemically stable and poorly water-soluble.As a result of their persistence and lipophilic properties, they are slowlydegraded in the environment and are bioaccumulated within the food chain.However, depending on the degree and the position of chlorination, theindividual PCDDs and PCDFs congeners possess different physical andchemical properties, which determine their environmental behaviour. Thecarry-over of dioxins from feed to food depends on different factors:

• Degree and position of chlorination of PCDDs/PCDFs: In the food chain,only the 17 congeners with 2,3,7,8-substitution are bioaccumulated (tetra-to octachlorinated PCDDs/PCDFs). These congeners are more stableagainst the main enzymes involved in the metabolism of aromaticcompounds in animals. Even within the group of the 17 congeners with2,3,7,8-substitution, a wide range of different transfer rates can beobserved. Whereas 2,3,7,8-TCDD, 1,2,3,7,8-PeCDD and 2,3,4,7,8-PeCDF have a relatively high transfer rate from feed to cow’s milk,2,3,7,8-TCDF, OCDD and OCDF have a low transfer rate. Othercongeners are between these extremes.

• Animal species: whereas in mammals predominantly 2,3,7,8-substitutedcongeners are found, fish can also contain non-2,3,7,8-substitutedcongeners. Birds and eggs normally contain mainly 2,3,7,8-substitutedcongeners, but in case of an actual contamination, also non-2,3,7,8-substituted congeners can be determined.

• Kind of the matrix: soil or fly ash adsorb dioxins much stronger thangrass or other digestible feedingstuffs, which reduces their bioavailability.

At steady-state, the amount of dioxins absorbed is equal to the amount ofdioxins which are eliminated. The time to reach steady state is quite long (inthe range of months). The half-life of individual congeners varies and can beassumed for excretion in milk for TCDD with about 40 days and for OCDDwith about 80 days.

6.2. Calculation of transfer factors

The numerous ways of expressing contaminant carry-over found in theliterature lead to considerable difficulties in interpreting different studies,and some standardisation is necessary. Therefore, the specific definitions ofdifferent authors such as bioconcentration factor, carry-over ratio, transferefficiency or transfer rates have to be used.

64

Factors for the transfer from feed to cow’s milk can be calculated accordingto the following formula:

transfer factorcc

pf

mf

f

mf= *

with

cmf = concentration in milk fat (pg/g),

cf = concentration in feed (pg/g),

pmf = daily production of milk fat (g),

f = daily amount of feed intake (g)

Transfer factors (TF) are dimensionless, whereas transfer rates (TR, alsocalled carry-over ratios, COR) can be calculated also as percentage of thedaily dose that is transferred to milk fat at steady state (TR = TF * 100).

6.3. Results from relevant investigations

The transfer from grass to milk (see table 12) was studied in detail byMcLachlan et al. (1990). There, the “PCDDs/PCDFs fluxes out of the cowinto milk” were calculated as “% input” and are equal to “transfer rates”. In astudy of Schuler et al (1997b) with feeding of grass over a period of twoyears with 4 sampling dates, for example a range between 0.06 and 0.7 wasobserved for 2,3,7,8-TCDD, with an average of 0.3.

Fürst et al. (1993b) investigated PCDDs/PCDFs levels in cow’s milk inrelation to their levels in grass and soil. The results indicated that thepathway air grass cow is more important than the pathway soil grass cow. Moreover, it became apparent that the carry-over forPCDDs/PCDFs between grass and cow’s milk differs significantlydepending on congeners. While 2,3,7,8-TCDD showed the highest carry-over, OCDD was accumulated less by a factor of almost 40 compared to2,3,7,8 TCDD. The relative ratio of the factors determined for each congenerconcerned was very similar irrespective of the dioxin levels in the grass atthe different sampling sites.

After administration of a single dose of contaminated fly ash to two cows,the bioavailability (as percentage) was low: 0.17 % resp. 1.21 % for TCDD(Slob et al., 1995). However, the bioavailability was comparable to theresults for grass of McLachlan et al (1990) and Schuler et al (1997b), whengrass of the neighbourhood of a large Municipal Solid Waste Incinerator wasfed to cows.

65

Table 12 Transfer factors for PCDDs/PCDFs from grass or citrus pulp tomilk (reported in literature)

McLachlan etal. (1990)

Slob et al.(1995)

Schuler et al.(1997b)

Malisch(2000)

loc. A loc. B Sep 94 May 95 June 95 Oct 95 mean

I-TEQ 0.20 0.40WHO-TEQ 0.442,3,7,8- TCDD 0.35 0.15 0.15 0.1 0.06 0.7 0.4 0.3 0.581,2,3,7,8- PeCDD 0.33 0.11 0.10 0.08 0.2 0.3 0.2 0.2 0.491,2,3,4,7,8- HxCDD 0.17 0.057 0.055 0.05 0.06 0.07 0.1 0.081,2,3,6,7,8- HxCDD 0.14 0.062 0.0651,2,3,7,8,9- HxCDD 0.18 0.031 0.0321,2,3,4,6,7,8- HpCDD 0.03 0.0062 0.0062 nd 0.03 nd 0.01 0.02

OCDD 0.04 0.0012 0.0008 0.004 0.02 0.001 0.009 0.0082,3,7,8- TCDF 0.07 0.0087 0.0078 0.01 0.02 0.04 0.02 0.02 0.0281,2,3,7,8- PeCDF 0.06 0.0040 0.0053 0.02 0.04 0.04 0.05 0.04 0.0382,3,4,7,8- PeCDF 0.25 0.12 0.12 nd 0.7 nd 0.2 0.5 0.581,2,3,4,7,8- HxCDF 0.19 0.043 0.043 0.05 0.1 0.1 0.04 0.07 0.331,2,3,6,7,8- HxCDF 0.16 0.036 0.035 0.302,3,4,6,7,8- HxCDF 0.14 0.042 0.042 0.191,2,3,7,8,9- HxCDF1,2,3,4,6,7,8- HpCDF 0.03 0.0039 0.0037 0.004 0.01 0.02 0.01 0.01 0.0311,2,3,4,7,8,9- HpCDF 0.08 0.0037 0.0064 0.042

OCDF 0.01 0.00 0.00 0.01 0.009 0.02 0.01 0.01 0.004

The same bioavailability as reported for grass was observed by Malisch(2000) for the transfer from contaminated citrus pulp to milk. Here, the useof highly contaminated lime was the source of the dioxin contamination ofthe citrus pulp. Lime is added to the wet peel for neutralization andconstitutes about 2 % of the dried citrus pulp. Lime is dissolved in acidconditions easily, thus PCDDs/PCDFs adsorbed on lime are much morebioavailable than e.g. PCDDs/PCDFs adsorbed on soil or fly ash. It can beassumed that lime is dissolved completely during the neutralization step ofthe citrus pulp and that PCDDs/PCDFs are adsorbed on the citrus pulp.Thus, the bioavailability of PCDDs/PCDFs in the contaminated citrus pulp ismore similar to grass with deposition of PCDDs/PCDFs from air. It wasconcluded that the absolute numbers of the transfer factors, at least of thepredominant congeners, were in line with the literature [Olling et al (1991),Jilg and Müller (1994), Tuinstra et al (1992) and Traag et al (1999)], andthat the general tendency (highest transfer rates for 2,3,7,8-TCDD, 1,2,3,7,8-PeCDD and 2,3,4,7,8-PeCDF; low transfer rates for higher chlorinatedPCDDs/PCDFs and 2,3,7,8-TCDF) was met.

A transfer rate for the I-TEQ-value and the WHO-TEQ-value (based onWHO TEFs) can be calculated only from specific experiments because theserates depend strongly on the congener pattern. These calculated values mustbe considered as purely indicative and cannot be generalized.

66

Concerning the PCDD/PCDF transfer from feed to meat, two studiesrevealed results that are somewhat contradictory. While Ruoff (1995) founda highly variable distribution of individual congeners between differenttissues of the same animal, a German-French joint programme, where foodsamples from different regions from the upper Rhine river valley werecompared, demonstrated, as to be expected, that the dioxin levels of differenttypes of meat, but not liver, are nearly the same when expressed on fat basis.

From a scientific point of view, comprehensive studies to calculate thetransfer of PCDDs/PCDFs from feed to meat would be extremely difficultand could face severe problems when compared to studies carried out todetermine the transfer of these compounds to milk or eggs. Different kind ofanimals (cows, pigs or poultry) would have to be studied to take into accountmetabolic specificities. As it may be anticipated that steady state is reachedafter 100 to 200 days, a large and statistically relevant number of animalswould have to be slaughtered, and different tissues, i.e. different types ofmuscles and fats, liver and kidneys, would have to be analysed to highlightdistribution specificities.

Previous studies have shown that free foraging chickens can take upPCDDs/PCDFs from soil and rapidly transfer them into eggs. Moreover,these studies indicated that the concentrations and congener profiles ofPCDDs/PCDFs in eggs of chickens appeared to be related to the soil onwhich they were raised. Thus, the official food control usually differentiatesdioxin levels in eggs depending on the type of housing, i.e. eggs from cagedchicken, foraging chicken, or from chicken kept on ground (Fürst et al,1993a). The transfer of PCDDs/PCDFs from soil into eggs of foragingchicken was studied by Stephens et al. (1990 and 1995) and Schuler et al(1997a).

Table 13 summarises the “transfer efficiencies” (calculated in g dry weight/glipids) proposed by the latter.

It appears that the transfer rates decreased with increasing chlorination of thecongeners over more than one order of magnitude from the PeCDD/PeCDFto OCDD/OCDF. The relatively moderate transfer efficiency observed for2,3,7,8-TCDD/TCDF could not be explained.

67

Table 13 Transfer efficiencies “soil to eggs” (calculated in g dry weight/glipids):

PCDDs/PCDFs Transfer2,3,7,8- TCDD 1.21,2,3,7,8- PeCDD 2.41,2,3,4,7,8- HxCDD 1.51,2,3,6,7,8- HxCDD 1.61,2,3,7,8,9- HxCDD 0.81,2,3,4,6,7,8- HpCDD 0.4

OCDD 0.12,3,7,8- TCDF 3.31,2,3,7,8- PeCDF 4.42,3,4,7,8- PeCDF 0.81,2,3,4,7,8- HxCDF 0.91,2,3,6,7,8- HxCDF 1.02,3,4,6,7,8- HxCDF 0.61,2,3,7,8,9- HxCDF 0.11,2,3,4,6,7,8- HpCDF 0.21,2,3,4,7,8,9- HpCDF 0.1

OCDF 0,1

Little is known about transfer of dioxins from feed to fish. Somebiota/sediment studies were carried out by Loonen et al. (1994), but thatstudy does not give information on dioxin transfer from feed to biota. As thedioxin and furan pattern in average fish oil and fish meal contains inparticular congeners which are readily bioavailable such as 2,3,7,8-TCDDand 2,3,7,8-TCDF, it is expected that a considerable amount of dioxins andfurans will be transferred from fish oil and fish meal to animals with a dietcontaining fish meal and/or fish oil.

A study on PCB transfer has been carried out by de Boer and Pieters (1991).In that study, it was shown that the transfer rates of PCBs from feed to eelwere ca 80% for the PCBs 52, 101, 153 and 180. These results correspondedwith those reported by Lieb et al. (1974) who found transfer rates of 68% forrainbow trout. This means that the major part of the PCBs is beingtransferred from the feed to the fish. Assuming a similar behaviour of orthoand non-ortho chlorobiphenyls and considering that PCB can determine ca80% of the total TEQ, at least 60% of the total dioxin + PCB TEQ in fishfeed is likely to be transferred to fish.

6.4. Use of TEQ values for carry-over calculation

The dioxin transfer from feed to food is essentially congener-specific.Calculated factors may vary considerably depending also on the matrixingested (e.g. soil, grass or fat) and the animal species. Therefore, taking intoaccount the very different congener patterns which are found in differentsamples, the determination of an unique transfer rate from feed to food is not

68

possible. Only in case of a consistent and well-defined congener pattern,overall transfer rates may be used. As a result, only a few publicationsmention a transfer factor calculated from the results of their specificexperimental conditions. As an example, the transfer rate from grass to milkcalculated on I-TEQ-base was in the range between 0.08 and 0.27 (Schuler etal, 1997b). For citrus pulp, Malisch (2000a) found a transfer factor of 0.40on I-TEQ-base and of 0.44 on WHO-TEQ-base. However these very limiteddata do not allow to propose a range of values for these factors. As aconsequence, it is not scientifically correct to use compiled dioxin data infeedingstuff based only on the TEQ-content, for calculation of a possibletransfer from feed to food.

7. ANIMAL HEALTH ASPECTS

A considerable amount of work on risk evaluation for the humans has beenpublished covering the many aspects of dioxin toxicity. In contrast, very little dataexist concerning the toxicity of dioxins for non-laboratory animals and farm animalsin particular. Until recently the toxic episodes described were related to the intake ofaccidentally contaminated feedingstuffs, but in these few cases no correlation hasbeen drawn between levels of real exposure of the animals (both for intensity andduration) and the severity of symptoms of disturbances suffered by the animals.

The alert that started the recent dioxin crisis in Belgium arose from an observationby breeders and egg producers of decreased laying (10 to 30%) and hatchability ofhen’s eggs, as well as intoxication signs (neurological syndromes) in the survivingchickens (Federal Ministry of Agriculture - Belgium, 2000). These eggs were laid byanimals exposed to feedingstuffs that were later confirmed to be contaminated bydioxins originating from PCB polluted animal fat. The highest dioxin concentrationmeasured in hen feedingstuffs was 781 ng I-TEQ/kg dry matter, which correspondsto an oral exposure of the birds of about 30 ng /kg live weight. The second episodeof feedingstuff contamination that occurred in Austria in 1999, due to the use ofcontaminated kaolinitic clay as a binder/anti-caking agent additive, did not give riseto toxic outbreaks in pigs and poultry. It must be noted that the feedingstuff dioxincontents were in a range of 4.8 to 6.2 ng I-TEQ/kg dry matter, that corresponds to anexposure of the animals of about 150 fold lower than during the Belgian event.

Considering worse case scenario using upper levels of contamination, thecontamination of the diet would reach 8 ng WHO-TEQ/kg diet (according to thehighest value obtained in the calculation for fishes in table B29). On the basis of adaily feed consumption of 2% of the body weight, the exposure of the animalswould be 0.006 and 0.16 ng/kg/day b.w. respectively, figures which are considerablylower than the exposure level that created toxic outbreaks in chickens reproduction.

Therefore, although no data is available on the impact of the diets, whencontaminated at the levels identified by SCAN, it may be anticipated that no adverseeffect would occur in mammals, birds and fish, provided that they are not challengedby severe accidental contaminations.

69

8. CONCLUSION

For humans, it is well known that more than 90% of the daily dioxin intake comesfrom food. In the same way it can be assumed that the animal dioxin body burdenderives mainly from feeding. Therefore feedingstuffs, and in some cases soil, are ofspecial concern as potential sources of dioxins.

Considering the sources of the contamination of feed materials by dioxins, PCBsand dioxin-like PCBs, the following points must be stressed:

• The ubiquitous environmental distribution of these compounds causes abackground contamination affecting all terrestrial plants directly grazed(pastures) or used as raw materials for animal feeding as well as theaquatic feed chain. This is also true for the soil that might contaminatefeed materials, or which can be directly ingested by the animals.

• In addition to background contamination, direct accidental pollution offeed materials may occur due to:

– the localised discharge of dioxins from industrial activities includingmaterials containing or generating such compounds, but also frompolluted waste water,

– the contamination of feed materials during their production (especiallythe use of contaminated chemical or chemical by-products),processing and transportation ,

– illegal practices or management failures during feed production

• Due to their physical (lipophilic nature) and chemical (stability)properties, these compounds accumulate in animal tissues and are alsotransferred to animal products (milk, eggs).

8.1. Identification of contaminated feed materials

Until the recent dioxin crises, only a very limited number of data wasavailable for dioxin contamination of most feed materials due to the lack ofawareness of the importance of feedingstuffs for the dioxin contamination offood of animal origin, but also for reasons of limited technical and financialresources.

The lack of sensitivity of a number of analytical techniques used, namely interms of upper bound limits of detection and determination, makes theinterpretation of the results difficult, with the risk of an overestimation ofbackground levels.

The frequency distribution of dioxin levels in feed materials shows that theaverage background contamination is in the range of 0.1 to 0.3 ng WHO-TEQ /kg DM (upper bound limit of determination) for feed materials ofvegetal origin and minerals, and 0.1 to 1 ng WHO-TEQ/kg fat for animal fat(fish oil excluded).

70

The high variability of the results may reflect local contamination conditionsthat most of the time cannot be clearly differentiated between backgroundand accidental contaminations. This is especially true for grass and foragesdioxin aerial contamination. However, on the basis of data supplied byMember States and open literature published data, the SCAN hasestablished, considering worst case assumptions, lower, mean and upperlevels of dioxins.

The ranking of the contamination levels indicates that:

(a) If fish oil and fish meal are used for diets, they can contributeconsiderably to the dioxin contamination of animal diets, even if theirpercentage in the diet composition is low. However, the origin of thefish is decisive, as fish oil and fish meal from Europe is much morecontaminated than the corresponding products from South Pacific(Chile, Peru)

(b) Animal fats can also contribute to the overall contamination due tothe lipophilic properties of dioxins and their bioaccumulation in thefood/feed chain.

(c) All the other and quantitatively major feed materials of plant(roughages, cereals, proteinaceous sources) and of animal (milk by-products and meat and bone meal) origin normally contain meanconcentrations around or below 0.2 ng WHO-TEQ/kg dry matter.Amongst these remaining feed materials, roughages present verywide concentration ranges depending on anthropomorphic genesisconditions and aerial dispersion of dioxins, which explains the uppervalues observed.

(d) Finally soil shows an even wider range of contamination, but itscontribution is limited to the consumption of the fraction adhering tofeed materials into contact, and to free range animal productions.Moreover, the bioavailability of the dioxins adsorbed onto the soil islow.

(e) Considering the rate of inclusion of feed materials combined withtheir impact (in percentage) on the total contamination of the diets,feed materials of fish origin (of European origin) and animal fatrepresent the major contributors to diets contamination. The impact isincreased when, in addition, such feed materials represent the mainquantitative component of the diets, like for instance for carnivorousfishes.

The contribution of dioxin-like PCBs to the contamination of feed materialsis not sufficiently investigated to be taken into account into the SCANassessment of the diet burden. However, the limited data available indicate acontribution similar to dioxins in all the feed materials, except for fish mealand fish oil where dioxin-like PCBs would increase the total TEQ valueabout 5 fold. This worsens the situation and should be taken intoconsideration for purposes of risk management.

71

8.2. Contribution to the contamination of food of animal origin

Depending on the degree and on the position of chlorination, the individualPCDDs and PCDFs show very different transfer rates. As a consequence, itis not scientifically correct to calculate transfer from feedingstuffs toproducts of animal animal origin on a TEQ basis only. The transfer can becalculated only on the basis of the congener profiles.

8.3. Impact on animal health

The only data available that allow us to relate farm animal exposure to toxicoutbreaks indicate that no adverse effect would occur in mammals, birds andfish if they are not challenged by severe accidental contaminations.

8.4. Need for data

There is a clear lack of (consistent) data in the feed sector. Most data havebeen issued after the different dioxin crises and the information they provideis poor in terms of sampling and geographical distribution. In addition, thedata available refer mainly (when not only) to dioxins. Data on dioxin-likePCBs are also needed for the evaluation of the diets contamination, as theycan significantly (multiplication by a factor of 5 in the case of feed materialsof fish origin, according to the available data on fish contamination) impacton the total TEQ calculation.

72

9. RECOMMENDATIONS

9.1. The reduction of human exposure to dioxins related to food consumption isimportant to ensure consumer protection. Food of animal origin is apredominant source of exposure of consumers to dioxins. As foodcontamination is directly related to feed contamination, consistent cross-sectorial actions must be taken to reduce final dioxin impact on humanhealth.

9.2. An integrated approach should be followed to reduce the dioxin incidence allalong the food chain, i.e. from feed materials to food producing animals thento humans. Taking measures on feed materials and feedingstuffs is thus animportant step to reduce the dioxin uptake by human.

9.3. As far as feed materials are concerned,

– Emphasis should be put on reducing the impact of the most contaminatedfeed materials, e.g. fish meal and fish oil from Europe, on overall dietcontamination. This could be achieved by substituting the mostcontaminated by lesser contaminated sources, by reducing their intrinsiccontamination or by using non (less) contaminated alternatives,continuing to meet the animal nutrient requirements.

– Any material (recycled products, raw materials, ingredients) used in themanufacturing of feed materials should be guaranteed for quality andsafety so that it would not become a source of contamination.

– Good manufacturing practices as well as use of Hazard Analysis andCritical Control Point (HACCP) principles should be introduced /continued at feed industry level to control the dioxin contamination alongthe different steps of the manufacture of feed.

– Efforts should be made to reduce dioxins contamination of feed resourcesat farm level (Good agricultural practices) and controlling localconditions of livestock production (e.g. direct environment of dairyfarms, free range animal production), in particular in areas where soilcontamination is elevated.

9.4. In addition, considering the impact of the environmental pollution on thecontamination of feed materials, measures implemented with the aim toreduce the general dioxin burden should be actively continued.

9.5. There is a need for data on dioxins, but also dioxin-like PCBs and PCBscontamination of the widest range of feed materials and feedingstuffs inorder to determine background levels so as to identify unknowncontamination sources. A particular effort should be put into obtaining moredata for feed materials for which a wide range of contamination is reported.

9.6. Scientific cooperation should be promoted in order to collect and collateinformation available in the different Member States at the EU level.

73

9.7. Monitoring programs could be organised at European level in order to widenthe current limited geographical basis of the information on feed materialscontamination.

9.8. Regarding the huge amounts of feedingstuffs being produced and sold worldwide, feed materials imported from third countries should be checked fortheir dioxin content with the aim to avoid additional dioxin burden onfeedingstuffs.

9.9. Data should be obtained from fully identified samples (contaminated or not)and techniques. In particular, two important aspects: indications or checks,whether samples were possibly contaminated, and the definition of the limitof determination have to be considered and reported to make data useful anduseable. The concept of upper bound determination limits should thereforebe implemented.

9.10. Analytical means should be developed and appropriate analyticalrequirements adopted according to the aim of the analysis carried out. Forinstance, in the case of checks of dioxins levels for control purposes, theanalytical limits of determination should be in the range of one fifth of theregulatory limits whereas for control of time trends of backgroundcontamination, the limit of determination should be clearly below the meanof the present background ranges for the different matrices.

9.11. Emphasis should be put also on quality and qualifications of laboratoriesinvolved in monitoring programs and control activities of or for feedproducers. Interlaboratory calibration studies should be promoted, usingreference materials with certified dioxin and dioxin-like PCBs contentswhich should be made available. With this regard, the Committee welcomesthe recent initiative of the European Commission (2000/C 290/05)10 invitingfor submission of proposals to support the development of CertifiedReference Materials, in particular in the field of environmental contaminantsin food and animal feed (topic IV.20) and specifically for PCDDs, PCDFsand PCBs in food and feed products. Such actions should be promoted.

9.12. Basic research is needed on studying the carry-over and establishingpertinent transfer factors for dioxin-like PCBs congeners from soil and feedto animals tissues and products (milk, eggs).

10 OJ n° C 290 of 13.10.2000, p. 4

74

10. BIBLIOGRAPHIC REFERENCES

ANON. (1999a).Fish meal and fish oil dioxin levels. International Fishmeal & Oil ManufacturersAssociation (IFOMA), St. Albans, UK.

ANON. (1999b).Overview of dioxins in Norwegian fish and marine products, Institute of Nutrition,Directorate of Fisheries (Annex 1).

ANON. (2000).Gesamtüberblick “Dioxin”. BFL, Institut für Futtermittel, Austria.

APPEL, K.E. (1989)Toxikologische Beurteilung polybromierter Dibenzodioxine und Dibenzofurane.Tätigkeitsbericht des Bundesgesundheitsamtes (BGA), Kap. 3.2.3

APPEL, K.E. (1992)Polybrominated Dibenzodioxins and Dibenzofurans - Toxicological Evaluation.The Toxicology Forum, Berlin, Germany, November 9-11, 1992, Report, p. 147 -159

AUDUNSSON, G.A. (2000).Email report from the Icelandic Fisheries Laboratories, Reykjavik, Iceland.

AUSTRIAN DATA (2000)Gesamtüberblick “Dioxin” BFL, Institut für Futtermittel, D.I. Franz Doppelreiter,12/01/00, 9p (not published)

BECHER, G., J.U. SKAARE, G.S. ERIKSEN and K. LUND LARSEN (1995).Dioksiner og dioksinelignende PCB I naeringsmidler I Norge. SNT-rapport nr. 9,Oslo, Norway.

BOER, J. DE, and PIETERS, H (1991).Dietary accumulation of polychlorinated biphenyls, chlorinated pesticides andmercury by cultivated eels, Anguilla anguilla L. Aquaculture and FisheriesManagement 22, 329-334.

BOER, J. DE, C.J.N. STRONCK, W.A. TRAAG and J. VAN DER MEER (1993).Non-ortho and mono-ortho substituted chlorobiphenyls and chlorinated dibenzo-p-dioxins and dibenzofurans in marine and freshwater fish and shellfish from TheNetherlands. Chemosphere 26, 1823-1842.

BREITBARTH, M.H., NOORDAM M.Y., ROEST, J.G. and VAN DER MENGELERS,M.J.B. (1999)Risico-inventarisatie van veevoedergrondstoffen en mogelijke beheersmaatregelen.Rikilt-dlo rapport 99.522.

CHANG, R.R., HAYWARD, D.G., GOLDMAN, L.R., HARNLY, M., FLATTERY, J.,and STEPHENS, R.O. (1989)Foraging farm animals as biomonitors for dioxin contamination. Chemosphere 19,481-486.

DÖRING, A. (2000)Monitoring zur Belastung von Einzelfuttermitteln und Zusatzstoffen mitpolychlorierten Dibenzodioxinen und -furanen. FEFAC, Brussels, Belgium.

75

EUROPEAN COMMISSION (1999a)Compilation of EU-Dioxin exposure and health data. Report produced for theEuropean Commission Directorate General on Environment, Task 2-Environmentallevels. Technical Annex, p.42

EUROPEAN COMMISSION (1999b)Compilation of data. on dioxin contamination of feedingstuffs made available byMember States delegation of the Standing Committee on feedingstuffs.Compilation established by the European Commission Directorate General onHealth and Consumer Protection.

EUROPEAN COMMISSION (1999c)Directorate General Health and Consumer Protection, DG(SANCO)/1198/1999-MR Final (08/10/99): Mission report on a routine mission carried out in Brazilfrom 05 to 09 July 1999, published on the internet:http://europa.eu.int/comm/dg24/health/fnaoi/reports/contaminants/brazil/fnaoi_rep_braz_1198-1999_en.pdf

EUROSTAT (1999)ISSN 1562-1340, Agriculture and Fisheries, theme 5-22/1999, 7/10/1999,Luxemburg

FEDERAL MINISTRY OF AGRICULTURE - BELGIUM (2000)Polychlorinated biphenyls (PCBs) and dioxins (PCDD/Fs) in the Belgiancontaminated food chain: sources, profiles and correlations. Final report preparedby F. Broeckaert and Alfred Bernard. March 2000.

FRIES, G.F., MARROW, G.S., and SNOW, P.A. (1982).Soil ingestion by dairy cattle. J. Dairy Sci, 65: 611-618

FÜRST, P, BECK, H., and THEELEN, R.M.C. (1992)Assessment of human intake of PCDDs and PCDFs from different environmentalsources. Toxic Substances Journal, 12, 133-150.

FÜRST, P., FÜRST, C., and WILMERS, K. (1993a)PCDD/PCDF in commercial chicken eggs-dependence on the type of housing.Organohalogen Compounds 13, 31-34

FÜRST, P., G.H.M. KRAUSE, D. HEIN,T. DELSCHEN AND K. WILMERS (1993b)PCDDs/PCDFs in cow’s milk in relation to their levels in grass and soil.Chemosphere 27: 1349 – 1357

GERMAN FEED MILLS (1999)Monitoring zur Belastung von Einzelfuttermitteln und Zusatzstoffen mitpolychlorierten Dibenzodioxinen und –furanen. Durchgeführt vom Bundesverbandder Mischfutterhersteller e.V., Fachverband der Futtermittelindustrie e.V. und derERGO Forschungsgesellschaft GmbH. Dec. 1999, 29 p. (not published)

GUOMUNDSSON, O. (1999) Letter to the European Commission Directorate General on Ariculture, 8 September1999, Feed, Seed and Fertilizer Inspectorate, Reykjavik, Iceland.

76

HÜLSTER, A. (1994)Transfer von polychlorierten Dibenzo-p-dioxinen und Dibenzofuranen(PCDD/PCDF) aus unterschiedlich stark belasteten Böden in Nahrungs- undFutterpflanzen. Thesis, University Hohenheim

HÜLSTER, A., MÜLLER, J.F., and MARSCHNER, H. (1994)Soil-plant transfer of polychlorinated dibenzo-p-dioxins and dibenzofurans tovegetables of the cucumber family (Cucurbitaceae). Environ. Sci. Technol. 28:1110- 1115

JILG, T. AND W. MÜLLER (1994)Zum carry-over von polychlorierten dibenzodioxinen (PCCD) und dibenzofuranen(PCDF) auf dem Belastungspfad Pflanze - Tier - Milch beim Rind. Tierärztl.Umschau 49: 169 - 172

JOBST, H., and ALDAG, R., (2000)Dioxine in Lagerstätten-Tonen. Z. Umweltchem. Ökotox. 12, 2-4.

KRAX, H. (1974)Geflügelproduktion, P. Parey Verlag, Hamburg, p. 84

KUBE, C; SCHÖPPE, G. (1998)Eintragspfade von organischen Schadstoffen über landwirtschaftliche Rohstoffe,Kraftfutter und Thermische Behandlungsschritte in Lebensmittel. UmweltmedizinForschung Praxis 3, 328-330

LAMPARSKI, L., NESTRICK, T.J. and STANGER, V.A. (1984).Presence of chlorodibenzodioxins in a sealed 1933 sample of dried municipalsewage sludge. Chemosphere, 13, 361-365.

LE DENMAT, M., DAGORN, J., AUMAITRE, A. AND VAUDELET, J.C. 1995Outdoor pig breeding in France. Pig News Information Vol 16 (N°1)13N-16N.

LIEB, A.J., BILLS, D.D., SINNHUBER, R.O. (1974).Accumulation of dietary polychlorinated biphenyls (Aroclor 1254) by rainbowtrout, Salmo gardneiri. J Agric. Food Chem. 22, 638-642.

LIEM, A.K.D., R.M.C. THEELEN, AND R. HOOGERBRUGGE (1996)Dioxinen en PCBs in voeding, resultaten van aanvullend onderzoek. RIVM report639102.005, Netherlands Institute of Public Health and the Environment,Bilthoven, The Netherlands.

LOONEN, H., J. R. PARSONS AND H.A.J. GOVERS (1994)Effect of sediment on the bioaccumulation of a complex mixture of polychlorinated-p-dioxins (PCDD) and polychlorinated dibenzofurans (PCDFs) by fish.Chemosphere (1994) 28 : 1433 - 1446.

LUNDEBYE-HALDORSEN, A.-K. AND Ø, LIE (1999)An overview of dioxins in Norwegian fish and marine products. NorwegianAgricultural inspection Service, Oslo, Norway.

MALISCH, R. (1996)Dioxin-like PCB in food and breast milk samples. Organohalogen Compounds 28:271-276.

77

MALISCH, R. (1998)Update of PCDD/PCDF - Intake from Food in Germany. Chemosphere 37: 1687 -1697

MALISCH, R. (2000a)Increase of the PCDDs/PCDFs-contamination of milk, butter and meat samples byuse of contaminated citrus pulp. Chemosphere 40: 1041-1053

MALISCH, R. (2000b)Update of PCDD/PCDF contamination of vegetable food. Summary of data ofresults of the Chemisches and Veterinäruntersuchungsamt Freiburg on the basis ofits annual reports.

MALISCH, R., FÜRST, P. (2000)Background contamination of feedstuff with dioxins Organohalogen compounds47: 322 - 325.

MARKS, H. (1985)Das Haushuhn. A. Ziemsen Verlag, Wittenberg

MAYLAND, H.F., FLORENCE, A.R., ROSENAU, R.C., LAZAR, V.A., TURNER,H.A. (1975)Soil ingestion by cattle on semiarid range as reflected by titanium analysis offaeces. J. Range Manage, 28:448.

MCKONE, T.E. (1994)Uncertainty and variability in human exposures to soil contaminants throughhomegrown food. A monte-carlo assessment. Risk Analysis 14, 449-462

MCLACHLAN, M.S., H. THOMA, M. REISSINGER AND O. HUTZINGER (1990)PCDDs/PCDFs in an Agricultural Food Chain, Part 1: PCDDs/PCDFs MassBalance of a Lactating Cow. Chemosphere 20: 1013 - 1020

MEHNER, A. (1968)Das Buch vom Huhn. E. Ulmer Verlag, Stuttgart

MONITORING SCHLESWIG-HOLSTEIN (1998)Dioxin Gehalte im Grundfutter - Auftrag des Ministers für ländl. Raüme desLandes Schleswig-Holstein

NEUMANN, G., HÜLSTER, A., RÖMHELD, V. (1999)PCDD/PCDF-Mobilizing Compounds in Root Exudates of Zucchini.Organohalogen Compounds 41:331-338

ÖBERG, L.G., ANDERSSON, R. and RAPPE, C. (1992)Organohalogen Compds. 9, 351-354.1992.

ÖBERG, L.G., WAGMAN, N., ANDERSSON, R. and RAPPE, C. (1994).Formation and degradation of polychlorinated dibenzo-p-dioxins, dibenzofuranesand biphenyls in compost. Organohalogen Compds. 18, 15-38.

OLLING, M., H.J.G.M. DERKS, P.L.M. BERENDE, A.K.D. LIEM AND A.P.J.M. DEJONG (1991)Toxicokinetics of eight 13C-labelled Polychlorinated Dibenzo-p-dioxins and -Furans in Lactating Cows. Chemosphere 23: 1377 - 1385

78

PETREAS, M.X., GOLDMAN, L.R., HAYWARD, D.G., CHANG, R.R., FLATTERY,J., WIESMULLER, T., STEPHENS, R.D., FRY, D.M., RAPPE, C., BERGEK, S.,HJIELT, M. (1991)Biotransfer and bioaccumulation of PCDD/PCDFs from soil: Controlled feedingstudies of chickens. Chemosphere 23, 1731-1741

RAPPE, C., OBERG, L. and ANDERSSON, R. (1999).Natural formation – a neglected source of polychlorinated dioxins anddibenzofurans. Organohalogen Compds. 43, 249-253.

RUOFF, K. (1995)Untersuchungen zum Übergang ausgewählter polychlorierter Dibenzo-para-dioxineund -furane nach oraler Supplementierung in die Milch laktierender Kühe.Dissertation, Universität Kiel

RUOFF, K., BLUTHGEN, A., UBBEN, E.H. (1999)Neuere Aspekte zur Kontamination der Milch mit polychlorierten Dibenzo-para-dioxinen und –furanen (PCDDs/PCDFs). Kieler MilchwirtschaftlicheForschungsberichte 51, 271-288

SCHÖPPE, G., KUBE-SCHWICKARDI, C. (1996)Untersuchungen zum Eintrag von PCDD und PCDF über Tierkraftfutter in diemenschliche Nahrungskette. UBA-FB 97-102

SCHÖPPE, G., KUBE-SCHWICKARDI, C. (1998)Eintragspfade von PCDD/PCDF-Belastungen über landwirtschaftliche Rohstoffeund Futtermittel in die menschliche Nahrungskette. Umweltwissenschaft undSchadstoffforschung 10, 107-108

SCHÖPPE, G., KUBE-SCHWICKARDI, C., BENDIG, H., REINDERS, G., SCHRAM,J., SCHULTE, M., TÜRK, J. (1997)Untersuchungen zum (Eintrag von polychlorierten Dibenzodioxinen (PCDD) undpolychlorierten Dibenzofuranen (PCDF) über Getreideprodukte- und Futtermittel indie menschliche Nahrungskette. Die Mühlen- und Mischfuttertechnik 134, 778-786

SCHULER, F., SCHMID, P. AND SCHLATTER, C. (1997a)The transfer of polychlorinated dibenzo-p-dioxins and dibenzofurans from soil intoeggs of foraging chicken. Chemosphere 34, 711-718

SCHULER, F., SCHMID, P. AND SCHLATTER, C. (1997b)Transfer of airborne polychlorinated dibenzo-p-dioxins and dibenzofurans intodairy milk. J. Agric. Food Chem. 45: 4162-4167

SIEVERS, S. and SCHACHT, U. (1994)Studies on dioxin formation in composting process under real conditions.Organohalogen Compounds. 18, 55-61.

SLOB, W., M. OLLING, H.J.G.M. DERKS AND A.P.J.M. DE JONG (1995)Congener-specific bioavailability of PCDDs/PCDFs and coplanar PCBs in cows:laboratory and field measurements. Chemosphere 31: 3827 - 3838

79

STEPHENS, R.D., M. HARNLY, D.G. HAYWARD, R.R. CHANG, J. FLATTERY,M.X. PETREAS AND L. GOLDMAN (1990)The bioaccumulation of dioxins in food animals II: Controlled exposure studies.Chemosphere 20: 1091 - 1096.

STEPHENS, R.D., M.X. PETREAS AND D.G. HAYWARD (1995)Biotransfer and accumulation of dioxins and furans from soil: chickens as modelfor foraging animals. The Science of the Total Environment 175: 253-273.

STEPHENS, R.D., PETREAS, M.X., HAYWARD, D.G. (1992)Biotransfer and bioaccumulation of dioxins and dibenzofurans from soil.Organohalogen Compounds 8, 377-380

STEPHENS, R.D., PETREAS, M.X., HAYWARD, D.G. (1995)Biotransfer and bioaccumulation of dioxins and furans from soil: chickens as amodel for foraging animals. Science of the Total Environment 175, 253-273

TACON, A.G.J. (1993)Feed ingredients for warmwater fish fishmeal and other processed feed stuffs. FAOFisheries Circular No. 856, Rome, Italy.

TRAAG, W., GLEADLE, A., HECHT, M., MAUSCH, R., TURRIO, A.,VERSTRAETE. F. (1998)Conclusions of the meeting of a group of experts “presence of dioxine in feedmaterials and the transfer rates of dioxins from feed in products of animal origin”,hold on 17 Nov. 1998; VI/8433/98

TRAAG, W.A., M.J.B. MENGELERS, C.A. KAN AND R. MALISCH (1999)Studies on the uptake and carry over of polychlorinated dibenzodioxins anddibenzofurans from contaminated citrus pulp pellets to cows milk. OrganohalogenCompounds 42: 201 - 204

TUINSTRA, L.G.M.TH., A.H. ROOS, P.L.M. BERENDE, J.A. VAN RHIJN, W.A.TRAAG AND M.J.B. MENGELERS (1992)Excretion of polychlorinated dibenzo-p-dioxins and -furans in milk of cows fed ondioxins in the dry Period. J. Agric. Food Chem 40: 1772 - 1776

VAN DEN BERG, MARTIN, ET AL. (23 AUTHORS) (1998)Toxic equivalency factors (TEFs) for PCBs, PCDDs, PCDFs for humans andwildlife. Environmental Health Perspectives 106: 775 - 792

VAN LEEUWEN, F.X.R., and YOUNES, M.M. (2000)Consultation on assessment of the health risk of dioxins: re-evaluation of thetolerable daily intake (TDI) : Executive summary. Food Additives andContaminants, Vol. 17, no. 4 (April 2000). pp. 223 - 240.

VARTIAINEN, T. AND A. HALLIKAINEN (1995)Polykloordibentso-p-dioksiinien (PCDDs/PCDFs) sekä PCB:n kertyminenkirjoloheen käytettäessä silakaa tai kuivarehua ravintona. Elintarvikevirastontutkimuksia 1/1995. (PCDDs/PCDFs and PCB accumulation into rainbow troutwhen fed either Baltic herring or dry fish feed).

80

WALKER, M.K., COOK, P.M., BATTERMAN, A.R., BUTTERWORTH, B.C.,BERINI, CH, LIBAL, J.J., HUFNAGLE, L.C. AND PETERSON, R.E. (1994)Translocation of 2,3,7,8-tetrachlorodibenzo-p-dioxin from adult female lake trout(Salvenilus namaycush) to oocytes: effects on early life stage development and sacfry survival. Can.J.Fish.Aquat.Sci. 51, 1410-1419.

81

11. ANNEX A: DATA AVAILABLE

11.1. Table A1

Dioxin results of feed materials and feedingstuffs samples (WHO-TEQ asupper bound determination limit of PCDDs/PCDFs determination) (Malisch,2000)

ng I-TEQ/kg dry matter

ng WHO-TEQ/kg dry matter

ng I-TEQ/kg fat

ng WHO-TEQ/kg fat

A1: forages and conservatesSamples without hints on sources for a possible dioxin contamination

No. of samples 67 67Mean 0.173 0.187Median 0.145 0.148Minimum 0.030 0.037Maximum 0.500 0.50990 %-percentile 0.311 0.33695 %-percentile 0.374 0.380

Contaminated samplesNo. of samples 6 6

Mean 5.900 6.604Median 4.082 4.146Minimum 0.832 0.835Maximum 20.155 24.07390 %-percentile 12.677 14.65495 %-percentile 16.416 19.364A2: tubers and rootsNo. of samples 2 2

Mean 0.021 0.036Minimum 0.016 0.030Maximum 0.026 0.043A3: cereals and seedsNo. of samples 8 8

Mean 0.035 0.037Median 0.008 0.010Minimum 0.005 0.007Maximum 0.182 0.18490 %-percentile 0.091 0.09295 %-percentile 0.137 0.138A4: by-products of plant origin

Samples without hints on sources for a possible dioxin contaminationNo. of samples 19 19

Mean 0.064 0.072Median 0.055 0.062Minimum 0.012 0.016Maximum 0.122 0.13290 %-percentile 0.120 0.12495 %-percentile 0.122 0.132

Contaminated samplesno. of samples 18 18Mean 7.317 7.925Median 7.295 7.886Minimum 4.635 5.131Maximum 10.145 11.20890 %-percentile 9.394 10.09395 %-percentile 9.681 10.290

82

ng I-TEQ/kg dry matter

ng WHO-TEQ/kg dry matter

ng I-TEQ/kg fat

ng WHO-TEQ/kg fat

Summary: all samples of group A without specific contaminationNo. of samples 96 96

Mean 0.137 0.150Median 0.110 0.121Minimum 0.005 0.007Maximum 0.500 0.509B1: feeds of animal origin - without pure fat or oilno. of samples 3 3 1Mean 0.100 0.108 0.34Minimum 0.055 0.061 0.34Maximum 0.145 0.160 0.34B2: pure fat or oil (definitely or probably of animal origin)No. of samples 19 19

Mean 0.93 1.04Median 0.63 0.68Minimum 0.24 0.26Maximum 5.68 6.6490 %-percentile 1.31 1.4295 %-percentile 2.14 2.42C: mineral feeds

Samples without hints on sources for a possible dioxin contaminationNo. of samples 22 22

Mean 0.113 0.126Median 0.064 0.071Minimum 0.013 0.015Maximum 0.454 0.48390 %-percentile 0.253 0.29895 %-percentile 0.330 0.355

Contaminatedno. of samples 30 30Mean 39.004 45.816Median 11.183 12.359Minimum 0.371 0.414Maximum 415.983 478.64690 %-percentile 64.806 73.83195 %-percentile 188.071 236.107D: compound feeds

Samples without hints on sources for a possible dioxin contaminationNo. of samples 33 33

Mean 0.061 0.068Median 0.051 0.052Minimum 0.008 0.011Maximum 0.232 0.24690 %-percentile 0.127 0.13695 %-percentile 0.138 0.155

ContaminatedNo. of samples 18 18

Mean 1.726 1.851Median 1.656 1.750Minimum 0.498 0.479Maximum 8.566 9.54690 %-percentile 2.045 2.13395 %-percentile 3.232 3.449

83

11.2. Table A2

PCDD/PCDF-content of vegetable food (in ng WHO-TEQ/kg dry weight)(Malisch, 2000)

Group 1group 1:potatoes

only

group 2all samples

besides zucchini

Group 2:Zucchini

only

group 3:all samplesbesides kale

group 3:kale only group 4 all

samples

Median 0.0683 0.0424 0.0395 0.2670 0.1354 0.2754 0.0777 0.1310Mean 0.1046 0.0571 0.0787 0.4335 0.1906 0.3601 0.1484 0.2008

Minimum 0.0143 0.0154 0.0170 0.1860 0.0235 0.0871 0.0194 0.014390 % percentile 0.2264 0.1089 0.1652 0.8696 0.3335 0.7766 0.310495 % percentile 0.3132 0.1494 0.2023 0.9559 0.4233 0.7802 0.4545

Maximum 0.4197 0.1899 0.2458 1.0051 0.9720 0.7806 0.7523 1.0051no. of samples 25 8 12 13 59 14 15 138

11.3. Table A3

Dioxin levels in fodder

FeedNumber of

samplesI-TEQ

(ng/kg dry matter)WHO-TEQ

(ng /kg dry matter) References

11 0.8 -Grass 29 1.6 -Not given 2.1 -Rye grass Not given 1.3 -Not given 1.0 -

15 0.13 -10 1.35 -Kale

5 0.48 -Silages 63 0.20 (0.12-0.5) -

EuropeanCommission, 1999a

Forages 67 0.17 (0.03-0.5) 0.2 (0.04-0.51) Malisch, 2000

11.4. Table A4

Dioxin levels in cereals and seeds

Feed Number ofsamples

I-TEQ(ng /kg product

or per kg dry matter)

WHO-TEQ(ng /kg product

or per kg dry matter)References

0.084 (upper bound)5 0.030 (lower bound) German feed mills 1999

2 0.16 (DM) Schöppe et al. 19971 0.009 (DM) 0.012 (DM) Malisch, 2000

NG 0.018 – 0.088 European Commission, 1999bRaw 0.071 (DM) Kube and Schöppe 1998

After sieving 0.090 (DM) Kube and Schöppe 1998Barley dust 0.330 (DM) Kube and Schöppe 1998

After cleaning 0.047 (DM) Kube and Schöppe 1998

Barley

NG 0.088 Austrian data 20000.080 (upper bound)4 0.027 (lower bound) German feed mills 1999

NG 0.005 - 0.150 European Commission, 1999b1 0.007 (DM) 0.009 (DM) Malisch, 2000

1 after drying 0.182 (DM) 0.184 (DM) Malisch, 2000

Corn

6 0.040 – 0.150 Austrian data 2000

84

Feed Number ofsamples

I-TEQ(ng /kg product

or per kg dry matter)

WHO-TEQ(ng /kg product

or per kg dry matter)References

NG 0.150 (DM) Schöppe and Kube-Schwickardi, 1996Oats NG 0.082 Austrian data 2000

0.093 (upper bound)6 0.045 (lower bound) German feed mills 1999

2 0.22 (DM) Schöppe and Kube-Schwickardi, 1996Rye

NG 0.084 Austrian data 2000NG 0.068 (DM) Schöppe and Kube-Schwickardi, 1996

Triticale NG 0.082 Austrian data 20000.084 (upper bound) German feed mills 199911 0.022 (lower bound)

13 0.140(0.0002- 0.390; DM) Schöppe and Kube-Schwickardi, 1996

4 0.019(0.005-0.052; DM)

0.021(0.007-0.053; DM) Malisch 2000

NG 0.007 – 0.086 European Commission, 1999bRaw 0.020 (DM) Kube and Schöppe 1998

After sieving 0.011 (DM) Kube and Schöppe 1998Wheat dust 0.440 (DM) Kube and Schöppe 1998

After cleaning 0.020 (DM) Kube and Schöppe 19985 0.010 – 0.087 Austrian data 2000

Wheat

NG + dust : 0.123 Austrian data 2000Peas NG 0.082 Austrian data 2000

NG 0.082 Austrian data 2000Soya beans 1 0.004 0.007 Malisch, 2000

8 0.035(0.004 – 0.182; DM) Malisch, 2000Cereals

and seeds NG 0.037(0.007 – 0.184; DM) Malisch, 2000

NG: not given

11.5. Table A5

Dioxin levels in by-products of milling

FeedNumber of

samples

I-TEQ(ng /kg product

or per kg dry matter)

WHO-TEQ(ng /kg product

or per kg dry matter)References

0.092 (upper bound)Oat bran 3 0.033 (lower bound) German feed mills 1999

0.091 (upper bound)4 0.034 (lower bound) German feed mills 1999

2 0.016(0.014-0.017; DM)

0.021(0.020-0.022; DM) Malisch 2000

NG 0.018 (DM) Schöppe et al. 1997NG 0.038 (DM) Kube and Schöppe 1998NG 0.040 – 0.100 European Commission, 1999b

Wheatbran

NG 0.082 – 0.100 Austrian data 2000

10 0.671(0.035-5.528 ; DM)Milling

By-products 9 0.153(without 5.528-

value; DM)

Ruoff et al. 1999

85

11.6. Table A6

Dioxin levels in by-products of starch industry

Feed Number ofsamples

I-TEQ(ng /kg product

or per kg dry matter)

WHO-TEQ(ng /kg product

or per kg dry matter)References

0.089 (upper bound)4 0.030 (lower bound) German feed mills 1999

3 0.130(0.089 - 0.190; DM) Schöppe and Kube-Schwickardi, 1996

1 0.103 (DM) 0.113 (DM) Malisch, 2000

Cornglutenmeal

2 0.152 – 0.185 Austrian data 2000Corn germ 0.086 (upper bound)

meal NG 0.023 (lower bound) German feed mills 1999

Manioc 2 0.380(0.050 – 0.710; DM) Schöppe and Kube-Schwickardi, 1996

11.7. Table A7

Dioxin levels in by-products of sugar industry

Feed Number ofsamples

I-TEQ(ng /kg product

or per kg dry matter)

WHO-TEQ(ng /kg product

or per kg dry matter)References

0.096 (upper bound)4 0.049 (lower bound) German feed mills 1999

NG 0.560 (DM) Schöppe et al. 1997

5 0.420(0.120-0.550; DM) European Commission, 1999b

1 0.073 (DM) 0.077 (DM) Malisch, 2000NG 0.220 – 0.340 Schöppe and Kube-Schwickardi, 1996

Sugar beetpulp

3 0.197 – 0.300 Austrian data 2000Molasses NG 0.116 – 0.150 Schöppe and Kube-Schwickardi, 1996Vinasse NG 0.082 Austrian data 2000

11.8. Table A8

Dioxin levels in by-products of brewery/distillery

Feed Number ofsamples

I-TEQ(ng /kg product

or per kg dry matter)

WHO-TEQ(ng /kg product

or per kg dry matter)References

0.074 (upper bound)1 0.016 (lower bound) German feed mills 1999

7 0.070(0.042-0.122; DM)

0.078(0.046-0.132; DM) Malisch, 2000Citrus pulp

NG 0.073(0.041 – 0.122; DM) Traag et al. 1998

Maltsprouts NG 0.233 Austrian data 2000

Yeast 1 0.106 (DM) 0.119 (DM) Malisch, 2000Brewers

grain 1 0.120 (DM) 0.122 (DM) Malisch, 2000

86

11.9. Table A9

Dioxin levels in by-products of oil industry and oils

FeedNumber of

samples

I-TEQ(ng /kg product

or per kg dry matter)

WHO-TEQ(ng /kg product

or per kg dry matter)References

0.088 (upper bound)CoconutsCake NG 0.044 (lower bound) German feed mills 1999

0.092 (upper bound)Line seedMeal 2 0.039 (lower bound) German feed mills 1999

0.084 (upper bound)4 0.030 (lower bound) German feed mills 1999

NG 0.069 (DM, Malaysia) Schöppe and Kube-Schwickardi1996

PalmKernelMeal

1 0.038 (DM) 0.040 (DM) Malisch, 20000.083 (upper bound)5 0.021 (lower bound) German feed mills 1999

NG 0.420 Schöppe et al. 19971 0.122 (DM) 0.132 (DM) Malisch, 2000

Rape SeedMeal

NG 0.002 Austrian data 2000Rape seed

Cake 1 0.014 (DM) 0.028 (DM) Malisch, 2000

2 0.049(0.012-0.086; DM)

0.056(0.016-0.096; DM) Malisch, 2000

0.082 (upper bound)7 0.030 (lower bound) German feed mills 1999

NG 0.420 (DM) Schöppe and Kube-Schwickardi,1998

Soya beanMeal

NG 0.030 Austrian data 2000Soya bean 0.105 (upper bound) German feed mills 1999

Shells NG 0.054 (lower bound)0.099 (upper bound) German feed mills 19994 0.042 (lower bound)Sunflower

Meal 2 0.04-0.084 Austrian data 2000

19 0.064(0.011 – 0.122) Malisch, 2000Byproducts

of plantorigin NG 0.072 (0.016 – 0.132) Malisch, 2000

87

11.10. Table A10

Dioxin levels in fish meal

Original data are presented in bold characters. Data in italics are calculatedon the basis of the original data, either by multiplying or by dividing them bya factor of 5 to correct for the PCBs concentration. (See Chapter 4.2.2.1)

WHO-TEQ (ng/kg fat)Material

Numberof

samples

I-TEQ(ng /kg fat) Dioxins only Dioxins and PCBs References

Fish meal Pacific 8 1.2 (0.18-2.1) 6 (0.9-10.5) Anon., 1999aFish meal Pacific 3 1.1 (0.6-1.5) 5.5 (3-7.5) Anon., 2000Fish meal/oil mixed, ofPacific and Europe origin 10 0.5 (0.08-6.8) 2.5 (0.4-34) Döring, 2000

Fish meal Europe 32 13.1 (1.0-47) 65.5 (5.0-235) Anon., 1999aFish meal Europe 5 3.7 (0.3-6.8) 18.5 (1.5-34) Anon., 2000

Fish meal Europe 6 8.3 (2.5-24) 41.5 (12.5-120)Lundebye-Haldorsen

and Lie, 1999

Fish meal Europe (North) 7 4.0 (1.0-11.3) 20 (5.0-56.5)Guomundsson, 1999,

Audunssnon, 2000

11.11. Table A11

Dioxin levels in meat and bone meal

Material Number ofsamples

I-TEQ(ng /kg product)

WHO-TEQ(ng /kg product) References

90 0.1-0.46 Industry2 0.2 Germany2 0.42 Germany

NG 0.26-0.73 Schöppe et al 1997NG 0.096-0.3 IndustryNG 0.07 FranceNG 0.17 IndustryNG 0.15 DenmarkNG 0.17 Germany

Meat and bonemeal

NG 0.1-0.75 Schöppe et al 1997NG 0.01-0.017 Industry

Blood meal 2 0.14 GermanyNG 0.25 Germany4 0.02-0.73 FranceFeather meal

NG 0.255 Industry

88

11.12. Table A12

Dioxin levels in milk replacers

Material Number ofsamples

I-TEQ(ng /kg DM)

WHO-TEQ(ng /kg DM) References

Dried whey NG 0.01 GermanyNG 0.424 ng/kg fat Industry

Refatted whey NG 0.02-0.09 IndustryMilk replacer 2 0.65 GermanyRefatted milk NG 0.5-0.74 Industry

11.13. Table A13

Dioxin levels in animal fat

Material Number ofsamples

I-TEQ(ng /kg product)

WHO-TEQ(ng /kg product) References

7 0.3 Austria2 0.225 Industry data

17 3.3 (0.56-10.4) Denmark5 1.08 Germany

NG 0.8 Fürst 1998NG 0.06-1.2 Germany3 0.62 Denmark2 1.7-4.0 France

NG 0.2-0.6 Malisch et al 1999NG 0.4-0.98 Ferrario et alNG 0.44 FranceNG 0.7 DenmarkNG 0.2-2.3 Malisch et al 1999NG 0.4-1.1 Schöppe et al 1997

Animal fat

NG 0.47 France3 0.66 Schöppe et al. 19977 0.87 Ruoff et al. 1999

13 0.95 Bosshammer, pers. Com20 2.02 Denmark37 2-3.3 Denmark17 3.3 Denmark

Mixed fat

5 1.1 Germany2 0.6-4.8 Kühn and Steeg 1993

NG 0.9 Winters et al. 19962 0.33-2.1 EU data

20 4.13 (0.33-30.8) Feil and Elis 1998NG 0.3-7.2 Malisch et al 19992 0.225 Germany

NG 0.57 France

Tallow

19 0.96 (0.54-20.2) Denmark

89

11.14. Table A14

Dioxin levels in fish oil

Original data are presented in bold characters. Data in italics are calculatedon the basis of the original data, either by multiplying or by dividing them bya factor of 5 to correct for the PCBs concentration. (see chapter 4.2.2.2)

WHO-TEQ (ng /kg fat)Material

Numberof

samples

I-TEQ(ng /kg fat) Dioxins only Dioxins and PCBs

References

Fish oil Pacific 6 0.95 (0.17-2.6) 4.75 (0.85-13) Anon., 1999aFish oil Pacific 5 0.2 (0.16-0.4) 1.2 (0.8-1.9) Liem et al., 1996Fish oil Europe 40 6.3 (1.0-20.1) 31.5 (5.0-100) Anon., 1999aFish oil Europe 2 3.2 (1.7-4.6) 16 (8.5-23) Anon., 2000

Fish oil Europe 15 3.5 (0.7-8.5)Nordic-TEQ 17.5 (3.5-42.5)

Becher et al., 1997+ unpubl. data

Fish oil Europe 6 3.0 (1.1-5.2) 15 (5.7-25.8) Liem et al., 1996

Fish oil Europe(North) 6 5.6 (1.7-8.4) 28 ( 8.3-41.8)

Vartiainen et al.,1995, + unpubl.

Data

Fish oil Europe(North) 17 5.2 (1.4-6.8) 26 (7.0-34)

Guomundsson,1999, Audunssnon,

2000

11.15. Table A15

Dioxin levels in soil

Summary of dioxin concentrations in soil from EU Member States (ng I-TEQ/kg d.m.)

Pasture Arable Rural Forest OthersAustria 1.6 – 14 <1 – 64Belgium 2.1 – 2.3 2.7 – 8.9Finland <1 – 30 <1 – 25 1 – 5 10 – 30Germany <2 – 45Greece <1 – 13 4.8 <1 – 8.6Ireland <1 – 43 1.9 – 3.1 <1Italy 1.4 6.0 1.8 – 20Luxembourg 2.2 – 16The Netherlands <1 – 8.4 <1 – 24.2Spain <1Sweden <1 – 87 >1 – 20United Kingdom

90

11.16. Table A16

Dioxin levels in binders, anticaking agents and coagulants

Material Number ofsamples

I-TEQ(ng /kg product)

WHO-TEQ(ng /kg product) References

2 Austria20 166 (1-1132) Germany3 0.63 (0.3-1.3) Germany

30 292 (5.3-1654) Germany11 286.6(51.2-509.3) Belgium3 32.8 (3.0-57.5) United Kingdom6 150 (76-298) France2 0.05-0.25 France4 0.062 France

20 87 (4.8-240) The Netherlands3 3.3-113 The Netherlands4 0.013-0.13 The Netherlands4 0.03-1.6 The Netherlands

NG 16.7 Denmark

Kaolinite

5 0.25 Austria1 <0.02 USA1 1.4 Germany1 1.7 The Netherlands

Montmorillonite,bentonite

1 <1.9 The NetherlandsZeolite 1 <4 The Netherlands

4 0.017 0.1 European Commmission, 1999aClinoptilolite NG <0.05 European Commmission, 1999a

NG 0.3 AustriaSteatite NG 0.02-0.47 Austria

4 0.29 The Netherlands5 0.23 IndustrySepiolite5 0.21 Industry

Lignosulfphonates NG 0.01-0.53 United KingdomSilicic acid,

colloidal silica NG 0.08-0.12 Anonymous

Sodiumaluminosilicate NG 0.25 Anonymous

91

11.17. Table A17

Dioxin levels in minerals / trace elements

Material Number ofsamples

I-TEQ(ng /kg product)

WHO-TEQ(ng /kg product) References

NG 0.3-0.37 AustriaNG 0.07 AustriaNG 0.3 GermanyNG 0.1 Austria2 0.3 Austria

NG 0.01 AustriaNG 0.08-0.2 European Commmission, 1999a

Phosphates

7 0.1-0.3 GermanyNG 0.049-0.082 AustriaNG 0.4 AustriaNG 0.19 The Netherlands

Chalk

NG 0.001-0.05 GermanyNG 0.05 Germany2 0.1 GermanySalt

NG 0.1 AustriaCa sulphate NG 0.19 The Netherlands

NG 0.15 Germany4 1.2-2.3 GermanyMg oxide

NG 0.27 AustriaMg Fe silicate NG 0.9 Germany

Zn oxide NG 0.02 Austria

92

12. ANNEX B: CALCULATIONS

In this annex, some abbreviations are used because of the size of the tables:

Ax, Bx, C: The codes A1 to A5, B1 to B3, and C are used in accordance withtheir definintion in chapter 3

Comp: composition of the feedingstuff considered, expressed inpercentage of dry matter

L: the calculation is based on the levels identified by SCAN as low foreach feed material composing the diet

M: the calculation is based on the levels identified by SCAN as meanfor each feed material composing the diet.

H: the calculation is based on the levels identified by SCAN as high,for each feed material composing the diet

Europe: feed materials originating from Europe

South Pacific: feed materials of fish origin (fish oil, fish meal) or concentratecontaining such feed materials.

The distinction has been made between Europe and South Pacific(Peru and Chile according to the literature) feed materials of fishorigin as literature shows difference in contamination by dioxins,which can impact on the total diet contamination, depending on thefeed materials used and their geographic origin.

In the tables, the total contamination of the diets, under SouthPacific, represents the sum of the value for the fish materials ofPacific origin (or in the case of ruminants for the concentrate IV)and of the values for the other diets components, which originatefrom Europe.

12.1. Ruminants

As most of the 20 examples of diets for ruminants involve mainly twodifferent components: roughage and concentrate used at different ratios, thecalculation has been done only with diets containing a different ratioroughage/concentrate.

Using a conservative approach, the Committee chose to use the levels ofcontamination identified for the Concentrate IV, for the calculation of thecontamination of the different ruminants diets, although more appropriateconcentrates may be used in practice depending on the type of productionand on the feed materials available.

93

12.1.1. Concentrates - Tables B1 & B2

Table B1 Estimation of the contamination of four different concentrates used in ruminants diets (expressed in ng WHO-TEQ/kg dry matter)

Concentrate n°I Concentrate n°II Concentrate n°III Concentrate n°IV

Comp Europe Comp Europe Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H % L M H % L M H L M H % L M H L M H

A3 Cereals & legumes 26 0.0026 0.0260 0.1040 55 0.0055 0.0550 0.2200 64 0.0064 0.0640 0.2560 62 0.0062 0.0620 0.2480

A4 By-products 70 0.0140 0.0700 0.4900 39 0.0078 0.0390 0.2730 25 0.0050 0.0250 0.1750 25 0.0050 0.0250 0.1750

B1 Fish meal - - 5 0.0020 0.0600 0.2800 0.0001 0.0004 0.0007 5 0.0020 0.0600 0.2800 0.0001 0.0004 0.0007

B2 Animal fat - 2 0.0100 0.0200 0.0660 2 0.0100 0.0200 0.0660 4 0.0200 0.0400 0.1320

C Premix 4 0,0008 0,0080 0,0200 4 0,0008 0,0080 0,0200 4 0,0008 0,0080 0,0200 4 0.0008 0.0080 0.0200

TOTAL 100 0,0174 0,1040 0,6140 100 0,0241 0,1220 0,5790 100 0,0242 0,1770 0,7970 0,0223 0,1174 0,5177 100 0.0340 0.195 0.8550 0.0321 0.1354 0.5757

Table B2 Estimation of the relative contribution of the different feed materials to the total contamination of four different concentrates usedin ruminants diets (expressed in percentage of the concentrate contamination). The percentages are calculated on the basis of tableB1.

Concentrate n°I Concentrate n°II Concentrate n°III Concentrate n°IV

Comp Europe Comp Europe Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H % L M H % L M H L M H % L M H L M H

A3 Cereals & legumes 26 15% 25% 17% 55 23% 45% 38% 64 26% 36% 32% 29% 55% 49% 62 18% 31% 29% 19% 46% 43%

A4 By-products 70 80% 67% 80% 39 32% 32% 48% 25 21% 14% 22% 22% 21% 34% 25 15% 13% 20% 16% 18% 30%

B1 Fish meal - - 5 8% 34% 35% 0% 0% 0% 5 6% 31% 34% 0% 0% 0%

B2 Animal fat - 2 42% 16% 11% 2 42% 11% 8% 45% 17% 13% 4 59% 21% 15% 62% 30% 23%

C Premix 4 5% 8% 3% 4 3% 7% 3% 4 3% 5% 3% 4% 7% 4% 4 2% 4% 2% 3% 6% 4%

TOTAL 100 100% 100% 100% 100 100% 100% 100% 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100% 100% 100%

94

12.1.2. Dairy cattle diets - Tables B3 to B6

Table B3 Estimation of the contamination (expressed in ng WHO-TEQ/kg drymatter) of two different diets used in ruminants reared for dairyproduction, with a milk yield of 5-10 kg/day (diets also applicable to drycows)

Ruminant diet n°1 Ruminant diet n°2

Comp Europe Comp Europe South Pacific

Feed materials % L M H % L M H L M H

A1 Roughage 100 0,1000 0,2000 6,6000 90 0,0900 0,1800 5,9400

Concentrate IV - 10 0,0030 0,0200 0,0860 0,0030 0,0140 0,0580

TOTAL 100 0,1000 0,2000 6,6000 100 0,0930 0,2000 6,0260 0,0930 0,1940 5,9980

Table B4 Estimation of the relative contribution of the different feed materials tothe total contamination of two different diets in ruminants reared fordairy production, with a milk yield of 5-10 kg/day (diets also applicableto dry cows). The percentages are calculated on the basis of table B3.

Ruminant diet n°1 Ruminant diet n°2

Comp Europe Comp Europe South Pacific

Feed materials % L M H % L M H L M H

A1 Roughage 100 100% 100% 100% 90 97% 90% 99% 97% 93% 99%

Concentrate IV - 10 3% 10% 1% 3% 7% 1%

TOTAL 100 100% 100% 100% 100 100% 100% 100% 100% 100% 100%

Table B5 Estimation of the contamination (expressed in ng WHO-TEQ/kg drymatter) of two different diets used in ruminants reared for milkproduction, with a milk yield of respectively 25-35 kg/day (diet n°4) and≥40kg/day (diet n°5)

Ruminant diet n°4 Ruminant diet n°5

Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H

A1 Roughage 60 0,0600 0,1200 3,9600 40 0,0400 0,0800 2,6400

Concentrate IV 40 0,0120 0,0800 0,3440 0,0120 0,0560 0,2320 60 0,0180 0,1200 0,5160 0,0180 0,0840 0,3480

TOTAL 100 0,0720 0,2000 4,3040 0,0720 0,1760 4,1920 100 0,0580 0,2000 3,1560 0,0580 0,1640 2,9880

Table B6 Estimation of the relative contribution of the different feed materials tothe total contamination of two different diets in ruminants reared for milkproduction, with a milk yield of respectively 25-35 kg/day (diet n°4) and≥40kg/day (diet n°5). The percentages are calculated on the basis of tableB5.

Ruminant diet n°4 Ruminant diet n°5

Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H

A1 Roughage 60 83% 60% 92% 83% 68% 94% 40 69% 40% 84% 69% 49% 88%

Concentrate IV 40 17% 40% 8% 17% 32% 6% 60 31% 60% 16% 31% 51% 12%

TOTAL 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100% 100% 100%

95

12.1.3. Beef cattle diets - Tables B7 to B10

Table B7 Estimation of the contamination (expressed in ng WHO-TEQ/kg drymatter) of two different diets used in ruminants reared for beefproduction, with an average daily gain of respectively 400-800 g/day(diet n° 6) and 800-1200 g/day (diet n° 7)

Ruminant diet n° 6 Ruminant diet n° 7

Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H

A1 Roughage 80 0,0800 0,1600 5,2800 75 0,0750 0,1500 4,9500

Concentrate IV 20 0,0060 0,0400 0,1720 0,0060 0,0280 0,1160 25 0,0075 0,0500 0,2150 0,0075 0,0350 0,1450

TOTAL 100 0,0860 0,2000 5,4520 0,0860 0,1880 5,3960 100 0,0825 0,2000 5,1650 0,0825 0,1850 5,0950

Table B8 Estimation of the relative contribution of the different feed materials tothe total contamination two different diets used in ruminants reared forbeef production, with an average daily gain of respectively 400-800 g/day(diet n° 6) and 800-1200 g/day (diet n° 7). The percentages are calculatedon the basis of table B7.

Ruminant diet n° 6 Ruminant diet n° 7

Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H

A1 Roughage 80 93% 80% 97% 93% 85% 98% 75 91% 75% 96% 91% 81% 97%

Concentrate IV 20 7% 20% 3% 7% 15% 2% 25 9% 25% 4% 9% 19% 3%

TOTAL 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100% 100% 100%

Table B9 Estimation of the contamination (expressed in ng WHO-TEQ/kg drymatter) two different diets used in ruminants reared for beef production:average daily gain of >1200 g/day (diet n° 8) and feed lots (diet n° 9).

Ruminant diet n° 8 Ruminant diet n° 9

Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H

A1 Roughage 70 0,0700 0,1400 4,6200 20 0,0200 0,0400 1,3200

Concentrate IV 30 0,0090 0,0600 0,2580 0,0090 0,0420 0,1740 80 0,0240 0,1600 0,6880 0,0240 0,1120 0,4640

TOTAL 100 0,0790 0,2000 4,8780 0,0790 0,1820 4,7940 100 0,0440 0,2000 2,0080 0,0440 0,1520 1,7840

Table B10 Estimation of the relative contribution of the different feed materials tothe total contamination two different diets used in ruminants reared forbeef production: average daily gain of >1200 g/day (diet n° 8) and feedlots (diet n° 9). The percentages are calculated on the basis of table B9.

Ruminant diet n° 8 Ruminant diet n° 9

Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H

A1 Roughage 70 89% 70% 95% 89% 77% 96% 20 45% 20% 66% 45% 26% 74%

Concentrate IV 30 11% 30% 5% 11% 23% 4% 80 55% 80% 34% 55% 74% 26%

TOTAL 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100% 100% 100%

96

12.1.4. Calves - Tables B11 & B12

Table B11 Estimation of the contamination (expressed in ng WHO-TEQ/kg drymatter) two different diets used in ruminants reared for beef production:average daily gain of >1200 g/day (diet n° 8) and feed lots (diet n° 9).

Ruminant diet n° 10

Comp Europe

Feed materials % L M H

A1 Roughage 10 0,0100 0,0200 0,6600

B5 Milk replacer 90 0,0540 0,1080 0,4320

TOTAL 100 0,0640 0,1280 1,0920

Table B12 Estimation of the relative contribution of the different feed materials tothe total contamination of two different diets in ruminants reared fordairy production, with a milk yield of 5-10 kg/day (diets also applicableto dry cows). The percentages are calculated on the basis of table B11.

Ruminant diet n° 10

Comp Europe

Feed materials % L M H

A1 Roughage 10 16% 16% 60%

B5 Milk replacer 90 84% 84% 40%

TOTAL 100 100% 100% 100%

97

12.2. Pigs

12.2.1. Piglets - Tables B13 & B14

Table B13 Estimation of the contamination (expressed in ng WHO-TEQ/kg dry matter) of three different diets used in young pigs betweenone and two months of age.

Pig diet n° 1 Pig diet n° 2 Pig diet n° 3

Comp Europe South Pacific Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H % L M H L M H

A3 Cereals & legumes 55 0.0055 0.0550 0.2200 59 0.0059 0.0590 0.2360 60 0.0060 0.0600 0.2400

A4 By-products 36 0.0072 0.0360 0.2520 24 0.0048 0.0240 0.1680 33 0.0066 0.0330 0.2310

B1 Fish meal 5 0.0020 0.0600 0.2800 0.0010 0.0070 0.0125 5 0.0020 0.0600 0.2800 0.0010 0.0070 0.0125 2 0.0008 0.0240 0.1120 0.0004 0.0028 0.0050

B2 Animal fat - 8 0.0400 0.0800 0.2640 2 0.0100 0.0200 0.0660

C Premix 4 0.0008 0.0080 0.0200 4 0.0008 0.0080 0.0200 3 0.0006 0.0060 0.0150

TOTAL 100 0.0155 0.1590 0.7720 0.0145 0.1070 0.5045 100 0.0535 0.2310 0.9680 0.0525 0.1780 0.7005 100 0.0240 0.1430 0.6640 0.0236 0.1218 0.5570

Table B14 Estimation of the relative contribution of the different feed materials to the total contamination of three different diets used inyoung pigs between one and two months of age (expressed in percentage of the total diet contamination). The percentages arecalculated on the basis of table B13

Pig diet n° 1 Pig diet n° 2 Pig diet n° 3

Comp Europe South Pacific Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H % L M H L M H

A3 Cereals & legumes 55 35% 35% 28% 38% 52% 44% 59 11% 26% 24% 11% 33% 33% 60 25% 42% 36% 25% 49% 43%

A4 By-products 36 47% 23% 33% 50% 34% 50% 24 9% 10% 17% 9% 13% 24% 33 28% 23% 35% 28% 27% 41%

B1 Fish meal 5 13% 37% 36% 7% 7% 2% 5 4% 26% 29% 2% 4% 2% 2 3% 17% 17% 2% 2% 1%

B2 Animal fat - 8 75% 35% 28% 76% 45% 38% 2 41% 14% 10% 42% 17% 12%

C Premix 4 5% 5% 3% 5% 7% 4% 4 1% 3% 2% 2% 5% 3% 3 3% 4% 2% 3% 5% 3%

TOTAL 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100% 100% 100%

98

12.2.2. Pigs for fattening – Tables B15 & B16

Table B15 Estimation of the contamination (expressed in ng WHO-TEQ/kg dry matter) of five different diets used in pigs for fattening.

Pig diet n° 4 Pig diet n° 5 Pig diet n° 6 Pig diet n° 7 Pig diet n° 8

Comp Europe Comp Europe Comp Europe Comp Europe Comp Europe

Feed materials % L M H % L M H % L M H % L M H % L M H

A3 Cereals & legumes 74 0.0074 0.0740 0.2960 40 0.0040 0.0400 0.1600 80 0.0080 0.0800 0.3200 69 0.0069 0.0690 0.2760 61 0.0061 0.0610 0.2440

A4 By-products 23 0.0046 0.0230 0.1610 53 0.0106 0.0530 0.3710 17 0.0034 0.0170 0.1190 23 0.0046 0.0230 0.1610 29 0.0058 0.0290 0.2030

B2 Animal fat - 4 0.0200 0.0400 0.1320 - 5 0.0250 0.0500 0.1650 4 0.0200 0.0400 0.1320

B2 Meat and bone meal - - - - 5 0.0050 0.0100 0.0250

C Premix 3 0.0006 0.0060 0.0150 3 0.0006 0.0060 0.0150 3 0.0006 0.0060 0.0150 3 0.0006 0.0060 0.0150 1 0.0002 0.0020 0.0050

TOTAL 100 0.0126 0.1030 0.4720 100 0.0352 0.1390 0.6780 100 0.0120 0.1030 0.4540 100 0.0371 0.1480 0.6170 100 0.0371 0.1420 0.6090

Table B16 Estimation of the relative contribution of the different feed materials to the total contamination of three different diets used inpigs for fattening (expressed in percentage of the total diet contamination). The percentages are calculated on the basis of tableB15.

Pig diet n° 4 Pig diet n° 5 Pig diet n° 6 Pig diet n° 7 Pig diet n° 8

Comp Europe Comp Europe Comp Europe Comp Europe Comp Europe

Feed materials % L M H % L M H % L M H % L M H % L M H

A3 Cereals & legumes 74 58% 72% 63% 40 11% 29% 24% 80 67% 77% 71% 69 19% 46% 45% 61 16% 44% 40%

A4 By-products 23 37% 22% 34% 53 30% 38% 55% 17 28% 17% 26% 23 12% 16% 26% 29 16% 20% 33%

B2 Animal fat - 4 57% 29% 19% - 5 67% 34% 27% 4 54% 28% 22%

B2 Meat and bone meal - - - - 5 13% 7% 4%

C Premix 3 5% 6% 3% 3 2% 4% 2% 3 5% 6% 3% 3 2% 4% 2% 1 1% 1% 1%

TOTAL 100 100% 100% 100% 100 100% 100% 100% 100 100% 100% 100% 100 100% 100% 100% 100 100% 100% 100%

99

12.2.3. Sows

12.2.3.1. Lactating sows – Tables B17 & B18

Table B17 Estimation of the contamination (expressed in ng WHO-TEQ/kg dry matter) of three different diets used in lactating sows.

Pig diet n° 9 Pig diet n° 10 Pig diet n° 11

Comp Europe South Pacific Comp Europe South Pacific Comp Europe

Feed materials % L M H L M H % L M H L M H % L M H

A3 Cereals & legumes 79 0.0079 0.0790 0.3160 58 0.0058 0.0580 0.2320 55 0.0055 0.0550 0.2200

A4 By-products 15 0.0030 0.0150 0.1050 31 0.0062 0.0310 0.2170 39 0.0078 0.0390 0.2730

B1 Fish meal 2 0.0008 0.0240 0.1120 0.0004 0.0028 0.0050 2 0.0008 0.0240 0.1120 0.0004 0.0028 0.0050 -

B2 Animal fat - 5 0.0250 0.0500 0.1650 3 0.0150 0.0300 0.0990

C Premix 4 0.0008 0.0080 0.0200 4 0.0008 0.0080 0.0200 3 0.0006 0.0060 0.0150

TOTAL 100 0.0125 0.1260 0.5530 0.0121 0.1048 0.4460 100 0.0386 0.1710 0.7460 0.0382 0.1498 0.6390 100 0.0289 0.1300 0.6070

Table B18 Estimation of the relative contribution of the different feed materials to the total contamination of three different diets used inlactating sows (expressed in percentage of the total diet contamination). The percentages are calculated on the basis of table B17.

Pig diet n° 9 Pig diet n° 10 Pig diet n° 11

Comp Europe South Pacific Comp Europe South Pacific Comp Europe

Feed materials % L M H L M H % L M H L M H % L M H

A3 Cereals & legumes 79 64% 63% 57% 65% 75% 71% 58 15% 34% 31% 15% 39% 36% 55 19% 42% 36%

A4 By-products 15 24% 12% 19% 25% 14% 24% 31 16% 18% 29% 16% 21% 34% 39 27% 30% 46%

B1 Fish meal 2 6% 19% 20% 3% 3% 1% 2 2% 14% 15% 1% 2% 1% -

B2 Animal fat - 5 65% 29% 22% 66% 33% 26% 3 52% 23% 16%

C Premix 4 6% 6% 4% 7% 8% 4% 4 2% 5% 3% 2% 5% 3% 3 2% 5% 2%

TOTAL 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100%

100

12.2.3.2. Pregnant sows – Tables B19 & B20

Table B19 Estimation of the contamination (expressed in ng WHO-TEQ/kg dry matter) of three different diets used in pregnant sows.

Pig diet n° 12 Pig diet n° 13 Pig diet n° 14

Comp Europe Comp Europe Comp Europe

Feed materials % L M H % L M H % L M H

A1 Roughage - - 32 0.0320 0.0640 2.1120

A3 Cereals & legumes 54 0.0054 0.0540 0.2160 29 0.0029 0.0290 0.1160 25 0.0025 0.0250 0.1000

A4 By-products 43 0.0086 0.0430 0.3010 64 0.0128 0.0640 0.4480 40 0.0080 0.0400 0.2800

B2 Animal fat - 5 0.0250 0.0500 0.1650 1 0.0050 0.0100 0.0330

C Premix 3 0.0006 0.0060 0.0150 2 0.0004 0.0040 0.0100 2 0.0004 0.0040 0.0100

TOTAL 100 0.0146 0.1030 0.5320 100 0.0411 0.1470 0.7390 100 0.0479 0.1430 2.5350

Table B20 Estimation of the relative contribution of the different feed materials to the total contamination of three different diets used inpregnant sows (expressed in percentage of the total diet contamination). The percentages are calculated on the basis of table B19.

Pig diet n° 12 Pig diet n° 13 Pig diet n° 14

Comp Europe Comp Europe Comp Europe

Feed materials % L M H % L M H % L M H

A1 Roughage - - 32 67% 45% 84%

A3 Cereals & legumes 54 37% 52% 41% 29 7% 20% 16% 25 5% 17% 4%

A4 By-products 43 59% 42% 56% 64 31% 43% 61% 40 17% 28% 11%

B2 Animal fat - 5 61% 34% 22% 1 10% 7% 1%

C Premix 3 4% 6% 3% 2 1% 3% 1% 2 1% 3% 0%

TOTAL 100 100% 100% 100% 100 100% 100% 100% 100 100% 100% 100%

101

12.3. Poultry

12.3.1. Broilers - Tables B21 & B22

Table B21 Estimation of the contamination (expressed in ng WHO-TEQ/kg dry matter) of three different diets used for broilers.Poultry diet n°1 Poultry diet n°2 Poultry diet n°3

Comp Europe South Pacific Comp Europe Comp Europe

Feed materials % L M H L M H % L M H % L M H

A3 Cereals & legumes 58 0.0058 0.0580 0.2320 70 0.0070 0.0700 0.2800 55 0.0055 0.0550 0.2200

A4 By-products 30 0.0060 0.0300 0.2100 18 0.0036 0.0180 0.1260 31 0.0062 0.0310 0.2170

B1 Fish meal 5 0.0020 0.0600 0.2800 0.0010 0.0070 0.0125 - -

B2 Animal fat 3 0.0150 0.0300 0.0990 5 0.0250 0.0500 0.1650 10 0.0500 0.1000 0.3300

B2 Meat and bone meal - 3 0.0030 0.0060 0.0150 -

C Premix 4 0.0008 0.0080 0.0200 4 0.0008 0.0080 0.0200 4 0.0008 0.0080 0.0200

TOTAL 100 0.0296 0.1860 0.8410 0.0286 0.1330 0.5735 100 0.0394 0.1520 0.6060 100 0.0625 0.1940 0.7870

Table B22 Estimation of the relative contribution of the different feed materials to the total contamination of three different diets used forbroilers (expressed in percentage of the total diet contamination). The percentages are calculated on the basis of table B21.

Poultry diet n°1 Poultry diet n°2 Poultry diet n°3

Comp Europe South Pacific Comp Europe Comp Europe

Feed materials % L M H L M H % L M H % L M H

A3 Cereals & legumes 58 20% 31% 28% 20% 44% 40% 70 18% 46% 46% 55 9% 28% 28%

A4 By-products 30 20% 16% 25% 21% 22% 37% 18 9% 12% 21% 31 10% 16% 27%

B1 Fish meal 5 7% 33% 33% 4% 5% 2% - -

B2 Animal fat 3 50% 16% 12% 52% 23% 17% 5 63% 33% 27% 10 80% 52% 42%

B2 Meat and bone meal - 3 8% 4% 3% -

C Premix 4 3% 4% 2% 3% 6% 4% 4 2% 5% 3% 4 1% 4% 3%

TOTAL 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100 100% 100% 100%

102

12.3.2. Layers – Tables B23 & B24

Table B23 Estimation of the contamination (expressed in ng WHO-TEQ/kg dry matter) of three different diets used for layers.Poultry diet n°4 Poultry diet n°5 Poultry diet n°6

Comp Europe South Pacific Comp Europe Comp Europe

Feed materials % L M H L M H % L M H % L M H

A1 Roughage 3 0.0030 0.0060 0.1980 3 0.0030 0.0060 0.1980 5 0.0050 0.0100 0.3300

A3 Cereals & legumes 70 0.0070 0.0700 0.2800 60 0.0060 0.0600 0.2400 60 0.0060 0.0600 0.2400

A4 By-products 15 0.0030 0.0150 0.1050 26 0.0052 0.0260 0.1820 24 0.0048 0.0240 0.1680

B1 Fish meal 3 0.0012 0.0360 0.1680 0.0006 0.0042 0.0075 - -

B2 Animal fat - - 2 0.0100 0.0200 0.0660

B2 Meat and bone meal - 3 0.0030 0.0060 0.0150 -

C Limestone 7 0.0070 0.0140 0.0350 6 0.0060 0.0120 0.0300 7 0.0070 0.0140 0.0350

C Premix 2 0.0004 0.0040 0.0100 2 0.0004 0.0040 0.0100 2 0.0004 0.0040 0.0100

TOTAL 100 0.0216 0.1450 0.7960 0.0210 0.1132 0.6355 100 0.0236 0.1140 0.6750 100 0.0332 0.1320 0.8490

Table B24 Estimation of the relative contribution of the different feed materials to the total contamination of three different diets used forlayers (expressed in percentage of the total diet contamination). The percentages are calculated on the basis of table B23.

Poultry diet n°4 Poultry diet n°5 Poultry diet n°6

Comp Europe South Pacific Comp Europe Comp Europe

Feed materials % L M H L M H % L M H % L M H

A1 Roughage 3 14% 4% 25% 14% 5% 31% 3 13% 5% 29% 5 15% 8% 39%

A3 Cereals & legumes 70 32% 48% 36% 34% 62% 44% 60 25% 52% 37% 60 18% 45% 28%

A4 By-products 15 14% 10% 13% 14% 13% 17% 26 22% 23% 27% 24 14% 18% 20%

B1 Fish meal 3 6% 25% 21% 3% 4% 1% - -

B2 Animal fat - - 2 31% 15% 8%

B2 Meat and bone meal - 3 13% 5% 2% -

C Limestone 7 32% 10% 4% 33% 12% 5% 6 25% 11% 4% 7 21% 11% 4%

C Premix 2 2% 3% 1% 2% 4% 2% 2 2% 4% 1% 2 1% 3% 1%

TOTAL 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100 100% 100% 100%

103

12.3.3. Turkeys – Tables B25 & B26

Table B25 Estimation of the contamination (expressed in ng WHO-TEQ/kg dry matter) of three different diets used for turkeys.

Poultry diet n°7 Poultry diet n°8 Poultry diet n°9

Comp Europe South Pacific Comp Europe Comp Europe

Feed materials % L M H L M H % L M H % L M H

A3 Cereals & legumes 47 0.0047 0.0470 0.1880 60 0.0060 0.0600 0.2400 55 0.0055 0.0550 0.2200

A4 By-products 45 0.0090 0.0450 0.3150 30 0.0060 0.0300 0.2100 34 0.0068 0.0340 0.2380

B1 Fish meal 3 0.0012 0.0360 0.1680 0.0006 0.0042 0.0075 - -

B2 Animal fat 2 0.0100 0.0200 0.0660 4 0.0200 0.0400 0.1320 8 0.0400 0.0800 0.2640

B2 Meat and bone meal - 3 0.0030 0.0060 0.0150 -

C Premix 3 0.0006 0.0060 0.0150 3 0.0006 0.0060 0.0150 3 0.0006 0.0060 0.0150

TOTAL 100 0.0255 0.1540 0.7520 0.0249 0.1222 0.5915 100 0.0356 0.1420 0.6120 100 0.0529 0.1750 0.7370

Table B26 Estimation of the relative contribution of the different feed materials to the total contamination of three different diets used forturkeys (expressed in percentage of the total diet contamination). The percentages are calculated on the basis of table B25.

Poultry diet n°7 Poultry diet n°8 Poultry diet n°9

Comp Europe South Pacific Comp Europe Comp Europe

Feed materials % L M H L M H % L M H % L M H

A3 Cereals & legumes 47 18% 31% 25% 19% 39% 32% 60 17% 43% 40% 55 10% 31% 30%

A4 By-products 45 35% 29% 42% 36% 37% 53% 30 17% 21% 34% 34 13% 19% 32%

B1 Fish meal 3 5% 23% 22% 2% 3% 1% - -

B2 Animal fat 2 39% 13% 9% 41% 16% 11% 4 56% 28% 22% 8 76% 47% 36%

B2 Meat and bone meal - 3 8% 4% 2% -

C Premix 3 3% 4% 2% 2% 5% 3% 3 2% 4% 2% 3 1% 3% 2%

TOTAL 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100 100% 100% 100%

104

12.4. Rabbits – Tables B27 & B28

Table B27 Estimation of the contamination (expressed in ng WHO-TEQ/kg dry matter) of three different diets used for rabbits. Two forrabbits for fattening (diets n° 1 and 2) and one for reproductive doe (diet n°3)

Rabbits diet n°1 Rabbits diet n°2 Rabbits diet n°3

Comp Europe Comp Europe Comp Europe

Feed materials % L M H % L M H % L M H

A1 Roughage 20 0.0200 0.0400 1.3200 18 0.0180 0.0360 1.1880 18 0.0180 0.0360 1.1880

A3 Cereals & legumes 25 0.0025 0.0250 0.1000 29 0.0029 0.0290 0.1160 33 0.0033 0.0330 0.1320

A4 By-products 50 0.0100 0.0500 0.3500 44 0.0088 0.0440 0.3080 44 0.0088 0.0440 0.3080

B2 Animal fat 1 0.0050 0.0100 0.0330 5 0.0250 0.0500 0.1650 1 0.0050 0.0100 0.0330

C Premix 4 0.0008 0.0080 0.0200 4 0.0008 0.0080 0.0200 4 0.0008 0.0080 0.0200

TOTAL 100 0.0383 0.1330 1.8230 100 0.0555 0.1670 1.7970 100 0.0359 0.1310 1.6810

Table B28 Estimation of the relative contribution of the different feed materials to the total contamination of three different diets used forrabbits (expressed in percentage of the total diet contamination). The percentages are calculated on the basis of table B27.

Rabbits diet n°1 Rabbits diet n°2 Rabbits diet n°3

Comp Europe Comp Europe Comp Europe

Feed materials % L M H % L M H % L M H

A1 Roughage 20 52% 30% 73% 18 33% 22% 67% 18 50% 27% 71%

A3 Cereals & legumes 25 7% 19% 5% 29 5% 17% 6% 33 9% 25% 8%

A4 By-products 50 26% 37% 19% 44 16% 26% 17% 44 25% 34% 18%

B2 Animal fat 1 13% 8% 2% 5 45% 30% 9% 1 14% 8% 2%

C Premix 4 2% 6% 1% 4 1% 5% 1% 4 2% 6% 1%

TOTAL 100 100% 100% 100% 100 100% 100% 100% 100 100% 100% 100%

105

12.5. Fish – Tables B29 & B30

Table B29 Estimation of the contamination (expressed in ng WHO-TEQ/kg dry matter) of the typical diets for omnivorous and carnivorousfish species.

Diet for omnivorous species Diet for carnivorous species

Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H

A3 Cereals & legumes 30 0.0030 0.0300 0.1200 11 0.0011 0.0110 0.0440

A4 By-products 56 0.0112 0.0560 0.3920 12 0.0024 0.0120 0.0840

B1 Fish meal 10 0.0040 0.1200 0.5600 0.0020 0.0140 0.0250 50 0.0200 0.6000 2.8000 0.0100 0.0700 0.1250

B1 Fish oil 2 0.0140 0.0960 0.4000 0.0032 0.0122 0.0520 25 0.1750 1.2000 5.0000 0.0400 0.1525 0.6500

C Premix 2 0.0004 0.0040 0.0100 2 0.0004 0.0040 0.0100

TOTAL 100 0.0326 0.3060 1.4820 0.0198 0.1162 0.5990 100 0.1989 1.8270 7.9380 0.0539 0.2495 0.9130

Table B30 Estimation of the relative contribution of the different feed materials to the total contamination of the typical diets foromnivorous and carnivorous fish species (expressed in percentage of the total diet contamination). The percentages arecalculated on the basis of table B29.

Diet for omnivorous species Diet for carnivorous species

Comp Europe South Pacific Comp Europe South Pacific

Feed materials % L M H L M H % L M H L M H

A3 Cereals & legumes 30 9% 10% 8% 15% 26% 20% 11 1% 1% 1% 2% 4% 5%

A4 By-products 56 34% 18% 26% 57% 48% 65% 12 1% 1% 1% 4% 5% 9%

B1 Fish meal 10 12% 39% 38% 10% 12% 4% 50 10% 33% 35% 19% 28% 14%

B1 Fish oil 2 44% 32% 27% 16% 11% 9% 25 88% 65% 63% 74% 61% 71%

C Premix 2 1% 1% 1% 2% 3% 2% 2 0% 0% 0% 1% 2% 1%

TOTAL 100 100% 100% 100% 100% 100% 100% 100 100% 100% 100% 100% 100% 100%